Iyonix woes

I’ve asked for help with this Iyonix a few times in the past but I still haven’t got it working. I was hoping to track down a TRM but still no joy with that holy grail!

The system continually freezes. Initially it was after quite some time and usually IDE activity that seemed to cause it.

I decided to revisit the issue, thinking I’d bypass IDE and use SCSI but although I got the Iyo to boot with the AKA32 and the 32bit rom, it won’t see any attached devices. During this effort, the freezing appears to have got worse. I’m now not so convinced it’s IDE that’s’ causing it.

Just booting, either clean or normal, I can get the Iyo to freeze by simply opening a window and resizing the window.

What I’ve tried;

Swapped the graphics card, swapped USB card, swapped the locations of the cards, completely vacated the 64bit slots (fitting jumper as required), ensured termination pcb snub is correctly aligned in PCI slot, swapped RAM, swapped PSU, rmkilled DMAManager, tried both IDE ports independently, tried different IDE disks (including SSD), jumper on PL601 is fitted.

I’m leaning towards a RAM issue and though I don’t know why, the DIMM socket looks a bit iffy. That is, at least one contact looks very slightly misaligned, i.e. sitting ever so slightly further back than the rest. Finding a DDR1 184pin through hole socket won’t be easy if that is the problem.

I’d like to be able to interact with the Iyo without the USB card installed but obviously that means no keyboard or mouse so no interaction! VNC server on RISC OS 5? Not sure how I’d get one installed at this point though.

So, anybody got any thoughts that might be helpful, apart from bin it :-) I won’t bother asking is anybody has a spare motherboard. I actually did get a replacement a few years back but the audio didn’t work on that.

If you got this far, thanks for reading.

I presume you’ve done the obvious and booted without the boot sequence to rule out any software?

My experience with the three Iyonix I have is they’re very fragile. I had to replace the USB card on one, the onboard Intel NIC failed on another which I replaced with a PCI Intel NIC (I’ve not tried reballing the chip yet because…) and the ROM Flash chip failed on two and had to be replaced.

My third is working although I avoid it for fear of it breaking, it only gets switched on when I absolutely have to test on an Iyonix.

My third is working although I avoid it for fear of it breaking,

I know the feeling :-)

I have clean booted, also installed a new, clean 5.30 !Boot. So I don’t think it’s software.

I know the Iyonix is known for being finicky and several components do fail at times but if describing those sorts of issues in a regular PC I would be getting suspicious of the power supply. If you are not getting a nice clean and stable power supply to the board and its components this could be the cause of most of these problems. If you have a suitable alternative power supply I would be trying this – ie swapping out the current power supply (note make sure the power supply is compatible with the requirements of the Iyonix).

Yeah, the PSU was one of the first things I tried. A closer, inspection of the RAM socket shows some anomalies but why I don’t know as I bought this Iyonix new from Castle back in the day.

So, the plot thickens. There is a crystal on the underside of the motherboard, opposite side of the NIC controller, marked ‘25.000 KD0117’. Each leg attached to separate smd capacitor.

This was a tad wobbly and on touching, one leg has detached from it’s capacitor. It can’t have been hanging on by much.

Likely candidate for freezing issue? During my testing over the last week or so, the board has been out of the case on the bench sat on thick towelling, likely not helping. Very professional :-)

I will attempt to re-attach and see what happens. eek!

The capacitor crystal might, just possibly, be the cause of your problems, but it’s unlikely.

Try running the Iyo without the NIC and see if it’s any better.

Other than that: cut your losses and buy a modern machine, at least a Raspberry Pi. That’s what I’ve done. The performance is much better, as is the reliability.

The capacitor might, just possibly, be the cause of your problems, but it’s unlikely.

I was thinking more the crystal but either way, I re-flowed the solder between the capacitor & crystal but still no joy.

I do have other hardware running RISC OS but you know how it is :-)

My apologies – I don’t know where my brain was when I typed “capacitor”, but it clearly wasn’t in proper control of my typing! Crystal is the word. I have appropriately edited my original post.

Each leg attached to separate smd capacitor.

Having things clagged on to the terminals of SMD caps probably isn’t the best for reliability, either. Ceramic caps are quite susceptible to thermal stress and cracking when reworked or left with blobs of solder that are bigger than specified by the manufacturer in their soldering guidelines, and after that can either go open or short circuit.

With no TRM or schematics how is one supposed to know how to repair these things? Testing the capacitors on either end of this crystal (with an Atlas ESR meter) gives ‘open circuit or low capacitance’ which means that the meter doesn’t recognise them as capacitors.

Anybody any clue about this crystal and/or the capacitors? It looks like a bodge job to me.

Edited to say there are no online images of the rear of these motherboards to compare, sadly.

Actually, I lied, there is a faulty one on fleabay with a pic of the rear showing the crystal & caps.

My apologies

No need to apologise. Disconnecting the nic was a good idea so I took out the cable, disabled Access & interface but still no joy.

Anybody any clue about this crystal and/or the capacitors? It looks like a bodge job to me.

I don’t have any specific information about this Iyonix crystal, but the most general crystal oscillator has a capacitor from each end of the crystal to ground. The value is likely to be around 47pF each, which is well below what any ordinary capacitance meter is likely to be able to measure – particularly if the capacitor is still in circuit.

Capacitors like that are orders of magnitude more reliable than quartz crystals, so they are highly unlikely to be your problem.

If you really want to pursue the crystal as the source of your problem, you can almost certainly replace it with any 25MHz crystal that you can get hold of. There’s a huge variety of cans that crystals can be bought in, so check what you’re getting.

If you really want to pursue the crystal as the source of your problem, you can almost certainly replace it with any 25MHz crystal that you can get hold of. There’s a huge variety of cans that crystals can be bought in, so check what you’re getting.

I don’t really want to mess if it’s not the issue, I guess I’m clutching at straws. The leg is now connected and I did used to have minor network issues in the past that maybe were unrelated to the freezing but are now cured with the re-flow. Maybe :-) Thanks for your help.

Was the Iyo reliable for years, and has gradually or suddenly become less reliable?

The fundamental problems you’re battling are typical of obsolete products, sadly. The service information is hard to come by; there’s not much current knowledge to crowd-source, because most Iyo owners have moved on (often because their Iyos, like yours, have become insufficiently reliable); the spares are hard and expensive to come by; and the Iyonix was always slightly iffy. Look how many people had to replace their PSUs and NICs because they were at the tolerance limit of working.

I abandoned mine years ago, first for a BeagleBoard, then for a BBxM; then when the BBxM died (a hard failure), I replaced it with a Raspberry Pi. It’s still right next to where I’m sitting, but I will eventually be forced to throw it out so as to reclaim the space. Even if, by magic, it started working again, I wouldn’t go back to it.

Anybody any clue about this crystal and/or the capacitors?

I asked John that very question yesterday as one of my Iyonix has that crystal and it also fell off. I asked him if there were any notes on the various mods/revisions Castle did. All three of my Iyonix are different for starters.

I suspect everything Castle wise has been lost to time, shame really as they’re not bad machines.

Anybody any clue about this crystal and/or the capacitors? It looks like a bodge job to me.

This mod was part of most (all?) supplied Iyonix boards. Given its location, it likely forms the 25MHz crystal oscillator for the 82540EM Gigabit Ethernet Controller, connected to K14 & J14 of the chip. The Design Guide for this series of chips shows 22pF load capacitors being used.
The Iyonix Schematic shows the 82540EM getting its clock source from U803 (ICS650-01 System Peripheral Clock Source) but I guess this didn’t work for some reason (noise due to long/poorly routed track?) so they had to cut that track (cut visible close to one of the load capacitors) and fit the crystal instead.

The Iyonix Schematic

You have schematics…please do share them!

You have schematics…please do share them!

The Iyonix TRM was sold as a separate item, much like the RiscPC TRMs, with a CD containing PDFs of the TRM and relevant datasheets for the chips used. As such, it will still come under copyright which I know people on here are hot on, preventing unauthorised distribution, unless Castle specifically forgo these rights.

This mod was part of most (all?) supplied Iyonix boards. Given its location, it likely forms the 25MHz crystal oscillator for the 82540EM Gigabit Ethernet Controller, connected to K14 & J14 of the chip. The Design Guide for this series of chips shows 22pF load capacitors being used.

Thank you for that info, very useful.

Having already reconnected the leg of the crystal to the capacitor I’m going to leave that for now and recap the board, starting with the ones I already have to hand. I’ve snipped all the 220uF 25v caps and their readings, out of the board, are as follows;

Cap#______uF_____ESR(Ohm)
_________________________
C717_____Open
C408_____200_____>20
C721_____188_____>20
C725_____193_____>20
C726_____212_____1.00
C409_____181_____>20
C411_____208_____9.7
C410_____211_____1.18
C724_____76______>20

I’ll replace them in the morning and see if there are any changes.

This mod was part of most (all?) supplied Iyonix boards. Given its location, it likely forms the 25MHz crystal oscillator for the 82540EM Gigabit Ethernet Controller, connected to K14 & J14 of the chipbeing used.

I’ve checked all three of mine, one board is rev.29 and has the crystal. The other two are rev.35 and don’t have it.

The rev.29 also has a capacitor mounted across two pins of the chip at U800.

It was the rev.29 that I put a PCI NIC into, so I suspect the crystal wasn’t mounted very well or simply didn’t fix the issue. I’ll have to check it again once I’ve got the flash chip replaced/reflashed.

Having fitted the new 220uF caps and a single 47ouF by the memory socket, the Iyo still freezes. It went for sone time though but ultimately failed. I’ll change the other caps but I’m not optimistic.

Any electronics guys comment on whether a 25MHz crystal with a load capacitance of 18pf and matching 18pF capacitors would likely be suitable for the network mod mentioned above?

@Jon, been meaning to ask, what NIC did you fit to your Iyonix?

Edited to also ask if you needed to do anything special to get it up and running?

It needs, iirc, the IntelPro 1000GT (or similar) PCI NIC, and then a new version of EtherK and the AutoSense code.