The Sinclair QL Forum

An internal microdrive cannot be directly connected to the MDV connector because, unlike the external microdrive, most of the motor control components components are on the QL motherboard. Of course, it would be possible to re-create these circuits on an extra PCB which could then be connected to the MDV connector.

Hi Keith

roninfourseven wrote:...would the interference prevent the drive from even being detected in slot 1? Or would it merely prevent it from operating correctly, because in y case the QL didn't even see the drive.

If, when you say "...the QL didn't even see the drive", did the (originally unit-1) drive spin-up at all when fitted to the right-hand slot when accessed with, say 'dir mdv2_', or did it appear totally unresponsive?

If not at all responsive in the right-hand slot - or even as I have observed on some failed units, causes the next drive in the chain to spin continuously even before/after being accessed - it could well point to the MDV ULA on the (originally unit-2) drive not correctly clocking-through the MD_SEL signal - only enough to trigger the QL mainboard's motor/LED for itself and its own internal circuitry, but not for anything further down the chain. A dodgy MDV ULA can play havoc with the shard MDV bus in all sorts of bizarre ways...

Equally, a border-edge working ULA might be more or less susceptible to the increased RF noise generated by the PAL encoder IC. If you don't use the RF modulator to generate your display, or otherwise do not need the composite PAL signal, you are quite safe to remove the PAL encoder IC altogether - even if just as a test.

Alternatively, the physical connection (fiddly at the best of times) between the drive board and the QL mainboard of one drive might not be secure after fitting in the other slot.

A failing tape-head is of course another possibility as you mention, but as the head circuitry sits 'behind' the MDV ULA, I wouldn't usually expect the head to behave differently in one slot versus in another (someone with more analogue electronics knowledge than I may have something more to say here.)

But otherwise, there is nothing in the MDV drive design that limits it to a given position in the 'chain'...

As I'm now a little confused myself by the references to unit-1 and unit-2 in your setup (before or after switching round? Or switched with another known-good drive from a different QL?), I'll spare you any more of my speculation for now - but what I would say is that having a spare, working _QL_ to switch drives between is another valuable 'test-tool' to identify good from border-line or completely-dead MDV units and, whilst it is possible to replace some of the head-board components, it's almost never worth the effort when compared to picking-up an external ZX Microdrive or another known-good bare QL drive - even my own (border-line perverse) level of fascination in this ageing technology doesn't stretch that far these day

martyn_hill wrote:Hi Keith

roninfourseven wrote:...would the interference prevent the drive from even being detected in slot 1? Or would it merely prevent it from operating correctly, because in y case the QL didn't even see the drive.

If, when you say "...the QL didn't even see the drive", did the (originally unit-1) drive spin-up at all when fitted to the right-hand slot when accessed with, say 'dir mdv2_', or did it appear totally unresponsive?

If not at all responsive in the right-hand slot - or even as I have observed on some failed units, causes the next drive in the chain to spin continuously even before/after being accessed - it could well point to the MDV ULA on the (originally unit-2) drive not correctly clocking-through the MD_SEL signal - only enough to trigger the QL mainboard's motor/LED for itself and its own internal circuitry, but not for anything further down the chain. A dodgy MDV ULA can play havoc with the shard MDV bus in all sorts of bizarre ways...

Equally, a border-edge working ULA might be more or less susceptible to the increased RF noise generated by the PAL encoder IC. If you don't use the RF modulator to generate your display, or otherwise do not need the composite PAL signal, you are quite safe to remove the PAL encoder IC altogether - even if just as a test.

Alternatively, the physical connection (fiddly at the best of times) between the drive board and the QL mainboard of one drive might not be secure after fitting in the other slot.

A failing tape-head is of course another possibility as you mention, but as the head circuitry sits 'behind' the MDV ULA, I wouldn't usually expect the head to behave differently in one slot versus in another (someone with more analogue electronics knowledge than I may have something more to say here.)

But otherwise, there is nothing in the MDV drive design that limits it to a given position in the 'chain'...

As I'm now a little confused myself by the references to unit-1 and unit-2 in your setup (before or after switching round? Or switched with another known-good drive from a different QL?), I'll spare you any more of my speculation for now - but what I would say is that having a spare, working _QL_ to switch drives between is another valuable 'test-tool' to identify good from border-line or completely-dead MDV units and, whilst it is possible to replace some of the head-board components, it's almost never worth the effort when compared to picking-up an external ZX Microdrive or another known-good bare QL drive - even my own (border-line perverse) level of fascination in this ageing technology doesn't stretch that far these day

Perhaps I can clarify. The microdrives have 1(left unit) & 2(right unit) marked on them as part of the black plastic moulding I showed in my earlier picture. The drive marked 1 sits in the left position of mdv1_ and is the one that has the shield over the ULA. The one marked 2 sits in the right position of mdv2_.

These are the original positions they came in, and both respond to commands, however only mdv2_ will read cartridges, mdv1_ will not.

I took both drives out and simply swapped them to try and get a working mdv1_. The drive that was originally mdv1_ became mdv2_, and the drive that was originally mdv2_ became mdv1_.

Just a straight swap. At this point the vMap was not fitted so could not interfere, as it was still in the post!

After swapping the drive positions neither drive worked. After returning them to their original positions, both worked as they had before.

To be clear, I don't have spare units to try as I only have one QL hence the straight swap. I did have another QL but foolishly sold it a couple of years back and now regret not having other spares to test with, like I do with nearly all of the other machines in my collection.

I'm confident the connections were ok (It took me a LONG time to get the damn things in under a magnifier) so I find it odd that they just didn't work when swapped based on what you've said.

TBH now I have a vDrive QL and have been able to use vMap to make the external vDrive mdv1_, the original mdv1_ is no longer a problem. I have used mdump_task with mdv2_ as a source to dump my small collection of cartridges to floppy and I can now use them off SD card after copying using Q-Emulator.

This is why I like retro though, this kind of stuff gives me a lot of pleasure, finding out and fixing etc.

Most of the QL<->Microdrive connectivity is simply a bus that should in no way be affected by the position where a drive is actually connected.

The only signal which is actually actively driven by connected drives is the COMMS IN/COMMS OUT line that is daisy-chained through the drive chain and actually could be affected by connected devices and their order in the chain. This signal is produced by the MDSEL signal of the 8302 and fed into the "shift register" made up by the chain of drives. (and also the signal that is modified to re-map the drives by the vMAP, I suppose).

It could be that your 8302 MDSEL output is particularily weak and your (original) mdv2 comms input generates a particularily high load on that line that would prevent this drive working directly as mdv1_ (when working as mdv1, this drive's COMMS IN is directly connected to the 8302, as mdv2 it connects to the COMMS OUT of the first drive)

It would be interesting to know whether physically swapping the drives would work with a vMap connected (because vMap re-generates MDSEL it could actually start to work in such a setup).

tofro wrote:Most of the QL<->Microdrive connectivity is simply a bus that should in no way be affected by the position where a drive is actually connected.

The only signal which is actually actively driven by connected drives is the COMMS IN/COMMS OUT line that is daisy-chained through the drive chain and actually could be affected by connected devices and their order in the chain. This signal is produced by the MDSEL signal of the 8302 and fed into the "shift register" made up by the chain of drives. (and also the signal that is modified to re-map the drives by the vMAP, I suppose).

It could be that your 8302 MDSEL output is particularily weak and your (original) mdv2 comms input generates a particularily high load on that line that would prevent this drive working directly as mdv1_ (when working as mdv1, this drive's COMMS IN is directly connected to the 8302, as mdv2 it connects to the COMMS OUT of the first drive)

It would be interesting to know whether physically swapping the drives would work with a vMap connected (because vMap re-generates MDSEL it could actually start to work in such a setup).

Yes, that's a very interesting thought regarding will it work when the vMap is used.

My main problem now is that I have everything reassembled and working, so I'm not over keen on disassembly!! I will give it some thought.

That's actually quite interesting and informative - I really just wanted to get a set up using the microdrive expansion slot to test internal mdv units to see if they were faulty or not without having to fit hem in the QL case. I have previously physically swapped mdv1 with mdv2 and both worked before and after they were displaced - so clearly as someone mentioned earlier both units can be swapped successfully without any problem, and the units are identical as far as I'm aware apart from the ULA shield (which can be swapped over anyway) - at least that is my experience

issue5 wrote:That's actually quite interesting and informative - I really just wanted to get a set up using the microdrive expansion slot to test internal mdv units to see if they were faulty or not without having to fit hem in the QL case. I have previously physically swapped mdv1 with mdv2 and both worked before and after they were displaced - so clearly as someone mentioned earlier both units can be swapped successfully without any problem, and the units are identical as far as I'm aware apart from the ULA shield (which can be swapped over anyway) - at least that is my experience

Yeah, it's weird why they don't work for me.

roninfourseven wrote:

issue5 wrote:That's actually quite interesting and informative - I really just wanted to get a set up using the microdrive expansion slot to test internal mdv units to see if they were faulty or not without having to fit hem in the QL case. I have previously physically swapped mdv1 with mdv2 and both worked before and after they were displaced - so clearly as someone mentioned earlier both units can be swapped successfully without any problem, and the units are identical as far as I'm aware apart from the ULA shield (which can be swapped over anyway) - at least that is my experience

Yeah, it's weird why they don't work for me.

Could it be bounded to your board version somehow?
I have the same problem. not tried yet to swap them each other