exxos's Atari, Amiga & retro forum

Ok, so I've responded to a few people on FB saying "I've just got an STE and I want to upgrade the RAM, please help", And I generally offer the advice to either go to the exxos store and buy a set from there, or trawl your way through eBay to find a set.

When I offer my advice, I also usually mention to be aware that you might have problems with parity RAM, as I did, which where really unreliable in both of my own STEs. I mention this because of what I've read, plus my own experiences.

I usually then get a slew of comments, some often rude, stating that parity RAM works just fine and why would I say otherwise? Well because it doesn't in my machines. Also why would you even bother with parity on a machine of the STs spec. The original machine did not use parity, why would you need it on the STE? I don't even have it in my MSTE.

Of course if it works for you great, but I mention it solely to help people make an informed choice.

So, because I've had this response quite a bit now I thought I'd put an open question out there: Parity or not?

@exxos I'm sure you had a section on this very subject, but I don't seem to be able to find it now.

It sounds like the "usual ignorance " of people..

I tested 100s of Simms a few years ago. Some I had for years , others brought more recently. I did post the info before can't find either..

Anyway, I don't remember exactly, but the first 40 sets all worked fine, but the 40 sets I found later did not. So I removed all the parity chips and they was fine then.

It really depends on brands of dram. Some don't care about floating lines, others do and cause the faults shown in my video. It is long known you shouldn't have floating signals, its like electronics 101. But people being people, don't care and as it works , won't listen. Who's to say the odd crash they could get now and then isnt down to parity Simms... They just wouldn't know..

Then they will get smart and run ram test to prove the point that the ONE set they have works fine..so falsely concluded every simm in the world I is fine... But its pot luck..

Even more complicated is even if RAM passes all day, it doesn't mean it is stable when ROM is being accessed.. Look at my video, doing ram test fails and corrupts the screen, then on desktop its fine..

Try and explain to people all this people.. 99% can't get past that ram can still cause problems and still pass RAM test fine all day.. Then they start to flame you so you go setup your own forum to get away from such people.. :lol: :roll:

As I said in the moderators forum, its not our job to teach people basic electronics , help people if they are willing to learn and listen,yes. otherwise hit the block button and just let um get on with it all..


EDIT:
The ram test software itself may or may not be reliable .. I will never forget when I tried memtest87 on my PC years ago.. Said RAM was fine.. I wasn't convinced.. I went back to memtest86 and it found faults right away..

EDIT2:

People wrongly assume just because the IO pins are not connected to anything that it doesn't matter.. But that's simply wrong. The RAM is still connected to the power rails. Its inputs left floating can oscillate causing increased power draw from the rails which leads to other problems. Its well known that the STE has various timing problems. Some are down to lack of proper GND layer. It doesn't take much to cause the machine to become unstable and parity simms *can* and *do* do exactly that. Again how problematic depends on the IC itself, the manufacture etc. So not all parity simms will have the problem.

I also noticed even apparent working parity simms can fail when routing the 32Mhz wire near them. I found this early on in my STE booster research and even later HERE. That problem is well known. Removing the parity simm alleviates those problems.

@rubber_jonnie , I see some of your posts on eBay and cringe thinking you should buckle your seatbelt. A lot of people are just spoiling for a fight which I’m usually up for but its really just a wasted tiring effort. Anymore I rarely post on eBay not to mention I visit rarely. It’s good you have all those ninja skills Jon.

Personally I'd prefer no parity, but the only SIMMs I had were 9 bits without discrete chips (only 3 blobs), this kind :
So I thought about putting a weak pull down on the STE side on the parity bit, but before I tested them as is and since they run great so far, I kept them in my STE. But I assume I'm just lucky and I keep in mind that if I have a problem later it might be the cause.

FB is... 'stupid'. The same question over and over. Hard to search and know how much info already exist there.

DoG wrote: 24 Jan 2021 14:47 FB is... 'stupid'. The same question over and over. Hard to search and know how much info already exist there.

I quite agree, FB is stupid, and especially so when you try to be helpful!!

PaulJ wrote: 24 Jan 2021 14:16 @rubber_jonnie , I see some of your posts on eBay and cringe thinking you should buckle your seatbelt. A lot of people are just spoiling for a fight which I’m usually up for but its really just a wasted tiring effort. Anymore I rarely post on eBay not to mention I visit rarely. It’s good you have all those ninja skills Jon.

I just like to be helpful, but I'm seriously considering avoiding these questions about RAM, and other similar questions that could be answered in5 minutes or less on Google. I am pretty patient though.

beel1 wrote: 24 Jan 2021 14:46 Personally I'd prefer no parity, but the only SIMMs I had were 9 bits without discrete chips (only 3 blobs), this kind :

1-MB-Simm-30-pin-60-ns-3-Chip-1Mx9-2.jpg

So I thought about putting a weak pull down on the STE side on the parity bit, but before I tested them as is and since they run great so far, I kept them in my STE. But I assume I'm just lucky and I keep in mind that if I have a problem later it might be the cause.

I have seen non parity 3 chips SIMMS, there were 2 3 bit and one 2 bit chips on there, which surprised me a little, not that I'm saying this is the rule, more like the exception.

Yours is a case in point too, it works just fine for you, yet both my STEs fail in a similar manner to the video exxos posted with parity SIMMS. Trouble is there is so much 30 pin SIMM stock out there, it is going to be pot luck if you happen to find a set with parity, as you have that works. And your experience proves you can :)

beel1 wrote: 24 Jan 2021 14:46 Personally I'd prefer no parity, but the only SIMMs I had were 9 bits without discrete chips (only 3 blobs), this kind :

1-MB-Simm-30-pin-60-ns-3-Chip-1Mx9-2.jpg

So I thought about putting a weak pull down on the STE side on the parity bit, but before I tested them as is and since they run great so far, I kept them in my STE. But I assume I'm just lucky and I keep in mind that if I have a problem later it might be the cause.

Might be a idea. Those style simms a couple sets went into my store stock, though they were fully tested out though.