My ZX81's stopped working. :-(

[1024MAK]

Good to hear that you have got your Zeddy back up and running

Mark

[stephen_usher]

Yeah, pity about the vLA81 though. I've no idea why it went phut.

There are no mods to the machine, not even composite video output so there shouldn't have been any way that it was overstressed electrically. I didn't even have the RAM pack installed at the time as I was just trying to tune the LCD TV into the signal.

[Bloodnok]

A bit late with the post but thought I'd add that occasionally new components do fail. I do test all main components such as CPLDs prior to assembly and fully test completed units before shipping but there will always be a small number of random failures soon after installation (anyone remember the bathtub curve?).

It is important that items are checked on delivery for any shipping damage and tested on the target hardware as soon as possible (i.e. not left on the shelf for 6 months I warranty items for 30 days as this would normally be plenty of time to weed out any premature failures. In the event an item cannot be resurrected I simply provide a replacement.

The most important thing is to contact me directly as soon as there is a problem - I'm probably your best bet for getting support on the hardware I design and make.

Charlie (vRetro Design)

[mk79]

A bit late with the post but thought I'd add that occasionally new components do fail. I do test all main components such as CPLDs prior to assembly and fully test completed units before shipping but there will always be a small number of random failures soon after installation (anyone remember the bathtub curve?).

I get the "test before shipping", but how do you test SMD CPLDs prior to assembly?

Cheers, Marcel

[stephen_usher]

A bit late with the post but thought I'd add that occasionally new components do fail. I do test all main components such as CPLDs prior to assembly and fully test completed units before shipping but there will always be a small number of random failures soon after installation (anyone remember the bathtub curve?).

It is important that items are checked on delivery for any shipping damage and tested on the target hardware as soon as possible (i.e. not left on the shelf for 6 months I warranty items for 30 days as this would normally be plenty of time to weed out any premature failures. In the event an item cannot be resurrected I simply provide a replacement.

The most important thing is to contact me directly as soon as there is a problem - I'm probably your best bet for getting support on the hardware I design and make.

Charlie (vRetro Design)

Don't worry, I wasn't blaming you or your QA. The vLA82 in the Spectrum is still happy (I think). T'was just "one of those things" which happens to hardware now and again.

[Bloodnok]

A bit late with the post but thought I'd add that occasionally new components do fail. I do test all main components such as CPLDs prior to assembly and fully test completed units before shipping but there will always be a small number of random failures soon after installation (anyone remember the bathtub curve?).

I get the "test before shipping", but how do you test SMD CPLDs prior to assembly?

Cheers, Marcel

Hi Marcel, I have a test board with a ZIF socket for the various CPLD types. I program each CPLD with a test program which in turn drives a set of multiplexed LEDs (8 x 8 matrix) attached to the IO pins. Any IO failure shows up in the resultant display as a stuck or missing pixel. If all OK then I flash the CPLD with the production firmware.

Cheers,

Charlie

[mk79]

Hi Marcel, I have a test board with a ZIF socket for the various CPLD types. I program each CPLD with a test program which in turn drives a set of multiplexed LEDs (8 x 8 matrix) attached to the IO pins. Any IO failure shows up in the resultant display as a stuck or missing pixel. If all OK then I flash the CPLD with the production firmware.

Wow, very professional, I'm impressed Do chips actually turn out to be defective out of the box? Never had this problem myself but then I only had pretty low volumes. Also, I tend to try to get my boards pre-assembled, so checking them beforehand becomes difficult

Cheers, Marcel

[Bloodnok]

Do chips actually turn out to be defective out of the box? Never had this problem myself but then I only had pretty low volumes. Also, I tend to try to get my boards pre-assembled, so checking them beforehand becomes difficult

Unfortunately, due to the global chip shortage, there are very few (or no) CPLDS available from the traditional, trusted, suppliers. This has forced me in recent months to source from other suppliers in what has become a 'wild west' market with exhorbitant prices and the accompanying risk of fake or relabled (black-topped) goods. In order to manage this I've had to resort to a more stringent procurement process as well as fully testing every CPLD in case I get sent a bad batch. I've already been reimbursed for an order of 100 which were obviously relabled (having proof of failure and being able to describe the testing process is invaluable in making a successful claim).

Any new designs I'm contemplating are with modern components but the major items are also practically unavailable from the usual sources (Lattice MACHX02 for instance). I certainly don't envy anyone coming into the retro design and build scene at this time - although the tools and technology are fantastic and accessible, trying to deliver anything is a daunting exercise on many levels. Luckily I enjoy a challenge

[mk79]

Do chips actually turn out to be defective out of the box? Never had this problem myself but then I only had pretty low volumes. Also, I tend to try to get my boards pre-assembled, so checking them beforehand becomes difficult

Unfortunately, due to the global chip shortage, there are very few (or no) CPLDS available from the traditional, trusted, suppliers.

My latest QL-SD variant is based on the XC95144XL and my usual source JLCPCB now wants like 30USD for a single chip (or 20 USD if I buy 100 or something like that). That used to be 8 or 9 USD. Curiously I found Mouser has them in stock at a quite reasonable price, usually it's the other way round We'll see, currently I'm not in the mood to redesign for a cheaper chip only to have that one become unavailable either or whatever

This has forced me in recent months to source from other suppliers in what has become a 'wild west' market with exhorbitant prices and the accompanying risk of fake or relabled (black-topped) goods. In order to manage this I've had to resort to a more stringent procurement process as well as fully testing every CPLD in case I get sent a bad batch. I've already been reimbursed for an order of 100 which were obviously relabled (having proof of failure and being able to describe the testing process is invaluable in making a successful claim).

Impressive, but also a lot of additional work, hopefully the market recovers sometime.

Any new designs I'm contemplating are with modern components but the major items are also practically unavailable from the usual sources (Lattice MACHX02 for instance). I certainly don't envy anyone coming into the retro design and build scene at this time - although the tools and technology are fantastic and accessible, trying to deliver anything is a daunting exercise on many levels. Luckily I enjoy a challenge

Well, the 5V compatibility drove me to the XC95144, but maybe it would have been better to just use lever shifters and a newer chip, no idea.

[Peter]

Well, the 5V compatibility drove me to the XC95144, but maybe it would have been better to just use lever shifters and a newer chip, no idea.

What made you move from the ispMACH 4000 family to XC9500?