Using ZX Microdrives on the QL
Posted: Wed Apr 05, 2017 8:01 am
I've wondered for a while why the original, standalone ZX Microdrives, refused to work with the QL. They usually read fine but, apart from the odd occasion, formatting and saving were out.
There are no significant mechanical differences as far as I can see, and for the read/write hardware the circuit is almost exactly the same. That is apart from R4/R5 in the ZX schematic and R34/R36 in the QL. The only reference to their function I’ve found is in the Interface 1 and Microdrive service manual which has them setting some sort of asymmetrical bias current (my interpretation) to overcome the effects of the DC erase.
Here’s the resistors I’m referring to:
In the ZX Microdrive these values are 150 and 180 ohms respectively while in the QL these are both set at 330 ohms.
Out of interest, I decided to try the QL values out in a ZX Microdrive and was surprised to find that it now worked just fine with the QL. Format performance was equivalent to the QL internal drives for the same cartridge. What’s more the drive worked fine on a Spectrum with no apparent side effects.
Intrigued by this, I then tried a range of values for R4/R5. These were both set to the same value for testing.
Before I could test effectively I needed to bypass the QL’s cartridge rejection feature. The QL guarantees a minimum of 100K (200 sectors) for a formatted cartridge. It does this by simply rejecting any that don’t make this minimum (that’s not quite how the sales literature put it). The Spectrum Interface 1 does not set any minimum and therefore it’s quite easy to see if the mechanism is not quite right, or a cartridge is getting past its use by date, as the formatted sector numbers reduce correspondingly.
For the QL, if the cartridge doesn’t make the 200 sectors then it simply fails. It also fails if sector 0 can’t be found as this is needed for the sector map. To overcome this in testing, I made a version of the Minerva rom with the minimum sector threshold removed. That way the QL would report on actual sectors formatted rather than just failing mysteriously.
I tested using the same cartridge (a known consistent performer) performing three passes for each resistor configuration and on each system. I then calculated the average formatted sector count.
This is what I found:
As can be seen, when connected to the Spectrum, and its lower data speed, just about any resistor value within the range of 150 ohms, to over 390 ohms, produced no statistically different result. When connected to the QL the acceptable range was more limited, as expected, but had a marked overlap with the values that also worked on the Spectrum. The original ZX asymmteric combination of 150/180 ohms failed to produce any usable sectors on the QL but worked fine of the Spectrum. Interestingly, setting R4/R5 to both the same low values, of 150 and then 180 ohms respectively, gave good results with 180 ohms being perfectly acceptable.
This lead to the question of why the choice of values for R4/R5 on the ZX Microdrive. Out of all the combinations I tried the only one that failed on the QL was the very one which was chosen for the ZX. Why would that be? Be great if someone could confirm what I’ve found.
In the meantime, try changing both R4 and R5 on a spare ZX Microdrive to 330 ohms and you may find that it now formats and saves OK on the QL.
Charles
There are no significant mechanical differences as far as I can see, and for the read/write hardware the circuit is almost exactly the same. That is apart from R4/R5 in the ZX schematic and R34/R36 in the QL. The only reference to their function I’ve found is in the Interface 1 and Microdrive service manual which has them setting some sort of asymmetrical bias current (my interpretation) to overcome the effects of the DC erase.
Here’s the resistors I’m referring to:
In the ZX Microdrive these values are 150 and 180 ohms respectively while in the QL these are both set at 330 ohms.
Out of interest, I decided to try the QL values out in a ZX Microdrive and was surprised to find that it now worked just fine with the QL. Format performance was equivalent to the QL internal drives for the same cartridge. What’s more the drive worked fine on a Spectrum with no apparent side effects.
Intrigued by this, I then tried a range of values for R4/R5. These were both set to the same value for testing.
Before I could test effectively I needed to bypass the QL’s cartridge rejection feature. The QL guarantees a minimum of 100K (200 sectors) for a formatted cartridge. It does this by simply rejecting any that don’t make this minimum (that’s not quite how the sales literature put it). The Spectrum Interface 1 does not set any minimum and therefore it’s quite easy to see if the mechanism is not quite right, or a cartridge is getting past its use by date, as the formatted sector numbers reduce correspondingly.
For the QL, if the cartridge doesn’t make the 200 sectors then it simply fails. It also fails if sector 0 can’t be found as this is needed for the sector map. To overcome this in testing, I made a version of the Minerva rom with the minimum sector threshold removed. That way the QL would report on actual sectors formatted rather than just failing mysteriously.
I tested using the same cartridge (a known consistent performer) performing three passes for each resistor configuration and on each system. I then calculated the average formatted sector count.
This is what I found:
As can be seen, when connected to the Spectrum, and its lower data speed, just about any resistor value within the range of 150 ohms, to over 390 ohms, produced no statistically different result. When connected to the QL the acceptable range was more limited, as expected, but had a marked overlap with the values that also worked on the Spectrum. The original ZX asymmteric combination of 150/180 ohms failed to produce any usable sectors on the QL but worked fine of the Spectrum. Interestingly, setting R4/R5 to both the same low values, of 150 and then 180 ohms respectively, gave good results with 180 ohms being perfectly acceptable.
This lead to the question of why the choice of values for R4/R5 on the ZX Microdrive. Out of all the combinations I tried the only one that failed on the QL was the very one which was chosen for the ZX. Why would that be? Be great if someone could confirm what I’ve found.
In the meantime, try changing both R4 and R5 on a spare ZX Microdrive to 330 ohms and you may find that it now formats and saves OK on the QL.
Charles