The Sinclair QL Forum

Hi,

I am hoping I can bring this topic up again (it has been discussed before) and I was hoping someone would comment/explain my observations.
I have tried multiple different microdrive units on various QL's and my experience always seems to be that they work well to read files but will fail to format, giving the error message: formatting failed.

Things I have checked:
I have made sure that the write protect tab is present.
I am using a microdrive cart that formats successfully on a known good drive.
I am using Charlie Inglie's power module with a 3 amp supply and the Tetroid power module.
A known working drive does format on this particular QL.

I understand that these drives are old and I have read other posts regarding looking at the voltage regulator or capacitors.
I also understand that the erase head is separate from the read/write head.

My question relates to the fact that after trying quite a few drives, the vast majority of them seem to work very well in terms of reading and it is the formatting that seems to be a very common issue.

Will the drive read reliably with ageing voltage regulator and capacitors and it is only the formatting that fails?
Could the Tetroid power module be introducing noise?

I understand that many of you have abandoned using the microdrives and I also realize that they were always unreliable but I am noticing that for me, the inability to format is the only function not working and it is the observed behaviour on multiple different drives. Surely all of them could not have failed in the same way? The issue can't be the QL itself because a know working drive does format on this particular machine.

Could anyone please provide comments, thoughts as to why almost all the drives I try will read perfectly but will fail to format?

Mark.

Hi Mark!

Others here will have something to add, I'm sure, but some factors you might consider are:

a) How QDOS/Minerva determines whether or not a format has failed. Minerva in particular will respond with a failure if the results of the format show either too high a Total Sector count, or too low Good Sector count (as compared to total found.) Minerva is programmed this way to reject 'marginal' cartridges/drives.

b) Drive speed: Drives that have had an adjustment to the motor speed can often push its performance just outside Format's acceptable parameters - set either too fast or too slow.

c) Writing as opposed to Formating: As well as reading, you may want to compare writing/saving new files in your testing to see if that also fails - the write process will accept a broader range of 'soft-errors' than Format as it will look for up to 4 errors against a given sector before giving up, whereas Format only looks twice - and expects both passes to match for a given sector before it determines that the sector is 'good'.

d) Erase head: Of course there are other hardware factors - many of which you've already looked in to - and the stability of the Erase (section of the) head in each drive could well be a factor which would affect both Write/Save and Format.

e) State of the ULA: The MDV ULA(s) could also be marginal - and possibly negatively impacting the Write amplification needed whilst still allowing the ability to Read.

f) As has been mentioned elsewhere, the alignment and pressure of the cartridge in the drive also can play a significant role in performance.

I myself have conducted pretty exhaustive testing across about 12 different drives-units of various standards over the years and seen such varying results from any given drive that it was difficult to formulate any real conclusions. Not that that stops me continuing to experiment...

I have built a 'breakout' adapter that allows me to measure the most significant signals using either of an analogue or digital scope without having to dismantle the MDVs or extract them from the case. This also allows me to add externally either an OPD or Spectrum drive, as well as an uncased QL MDV unit (with some extra hardware.)

Good hunting

Thank you,

I'd be very interested in your breakout adapter. Would it be hard for a beginner to build? I can solder and follow directions but I am only a beginner with regards to electronics.

I also wanted to post a follow-up observation: I was experimenting this morning and decided to sort my drives into "matching pairs" - that is to say that I simply matched them up by the appearance of their pcb, as I have an assortment of different aged drives.

I then installed a matching pair and now both appear to format!

Has anyone else observed that behaviour? - where two drives of different versions would not work together? I'm not sure if this was purely coincidence so I will keep experimenting but I am happy to have at least these two with the ability to format.

Like others have mentioned, it is a bit disconcerting that multiple attempts at formatting the same cart yield a different number of successful sectors - must have been scary to trust data to it, back in the day.

I am using the Minerva ROM.

I have also had the experience of going to bed delighted that a drive is now functioning, only to have it not work again the next day!

Like you, despite 6 months of playing with these drives and have not lost the incentive to keep trying I'd love to get my hands on a pair of Samsung drives but, living in Canada, the number of eBay sellers who will ship to me are very low and I'm also sure people don't sell them often.

Hi again Mark!

The breakout adapter is mostly just a LOT of soldering and crimping of MDV cables - happy to write-up something next week about how i designed and built it.

As for your interesting 'matching pair' observation - I myself am not aware of any interdependency or indeed any real difference between units fitted as MDV #1 and #2 (bar the added RF shield typically added around the MDV#1 ULA.) I'm guessing that was coincidence - all but the two now working units having some ULA fault.

What I have observed is that a failing MDV ULA can not only hit performance/functionality of the drive to which it is fitted, but also - due to the shared bus-topology of the MDV design - can knock-out any other connected MDV units. When I test a new drive, I typically remove any others to eliminate such interference...

I don't know if it would be of any help, but my MDI driver http://www.dilwyn.me.uk/utils/MDIdriver108.zip has a SuperBASIC program to read raw sectors from a cartridge, and store them in memory.

You could try to read a cartridge that has failed to format, and see what if anything can be read from it with the program. From memory, FORMAT makes one pass of the cartridge creating the headers and the sectors with a set bit pattern of data. Then makes another pass using a different data bit pattern.

I've never tried reading a cartridge that failed a format, but you might be able to determine if the format fail is on the first or second pass. Or the number of good/bad sectors.

Thank you all for your responses.

I shall try the microdrive driver.

I also wanted to clarify that I am aware that there is no difference between mdv1_ and mdv2_, what I had meant by pairing them was that I matched the drives by what appeared to be the same drive version/printed circuit board. I seem to have quite a few different versions, for example some have a resistor mod across the ULA chip and some have the diode mod on the front, others on the back of the pcb.

I'm sure it should not matter but I only had success after pairing two of the same drive version.

I realize that these drives are old and many people have had failures but it sill doesn't sit well with me that the QL's for most of us seem to be working, with their original capacitors and yet the microdrives are so fragile. Also, why would there be so many ULA failures and all those failures are affecting just the formatting?

I am using Minerva so I shall try the original ROM too, as the post above seems to suggest it might be less likely to reject a cart.

Just to satisfy my own curiosity, I will try another two "matching pairs" before giving up on my non-working drives.

I wish there was a version of vDrive (an incredible product especially when paired with vMAP) that could fit in the original drive area, rather than act as an external drive because my obsessive nature needs something working in that drive space