Internal ram upgrade to 640k - continued

[Dave]

tcat -

I apologise. I'm sorry your thread got hijacked for this. Please PM me.

[Dave]

I say "advanced technology" compared to anything in a QL from the 80s, with one exception that I had a hand in designing and manufacturing. We made about 8,000 SuperQBoards v3, and I bet most of them are still around in boxes and store-rooms around the World. It was the second best selling brand, beaten only by Miracle Systems.

Big picture: I am trying to create a whole line of expansions with no development budget. I have to plan carefully because each expansion exhausts the budget and has to replenish it. The felt pads and battery adaptors and SuperRAM have to fund the way for developing and kitting UltimIDE and UltraQ, and they in turn have to bring in sufficient funds to design *a whole computer*.

Sorry that my saving 2€ upsets you. It works. It works as well as your solution. 100 x 2€ is 200€ or $277. $277 is almost half the PCB production cost for UltimIDE.

It's a big deal for SuperRAM, which has very low margin. UltimIDE is a more complex and expensive product which will have a slightly bigger margin (in $ if not in %). UltraQ will be more expensive again and have slightly higher margins. It's a business reality. If you want to ever see QL2, this is how the money will be raised to develop and build and ship it.

If you want to start your own QL accessory business, I encourage you! You'll soon learn the business realities of working in a very small and uncertain market. For example, QL-SD could be a competitor to my UltimIDE (I'll probably lose a few sales to it, and they might lose a couple of sales to UltimIDE). However, instead of looking at it that way, we went out of our way to design a large RAM expansion that is uniquely suited to QL-SD and that will make it a much nicer product to use. By making QL-SD stronger, we help the whole QL community.

That's how we can have nice things.

If you'd like to start your own self-sufficient QL accessory business, and fund the development, manufacture and distribution of a first major product for your company, you'll quickly understand the choice I made.

Sticking a fancy regulator on SuperRAM won't make it better, faster, cheaper. It will make it draw a tiny bit less current and increase the price and complexity, while delaying delivery by a few days.

Mostly, I can't believe I have spent this amount of time arguing the point with you. If you think it should be done differently, you're welcome to start your own company and make them your way.

I'd even help you do it. It would benefit the community, which is why I am doing this in the first place. 'cos I'm such an ass.

[Dave]

Brane2,

You've said a lot of things I can agree with. I know your heart's in the right place. After I offered you a chance to specify the components on the board, with a couple of strings about them being better enough to justify the cost, you replied with:

But our "Decider" has on his family's coat of arms an ass as symbol of universal tool that got his ancestors through history so far, so he felt he owes to them to stick to tradition.
And he went for first option available.

I explained my decision isn't based on just one or two factors, but many competing factors, and that I am willing to change it if you could clearly show a better plan that wasn't too expensive.

I'd rather spend this time thinking of something that benefits the users, that they can see and feel, like mirrored video RAM and mirrored internal main memory. Like no-refresh SRAM that runs at full speed. Like how it will work with expansions.

If I worry about Every Little Detail No Matter How Small, it'll never get done. That's no use to anyone.

[vanpeebles]

It was meant as simply a question, not a critiqe or some personal "my kung fu is better than yours" attack or somesuch.

Snake fist style or eagles claw?

[1024MAK]

Excuse me, before we get a full scale bun fight, go have a look at this link.
Now of course you can buy the individual parts, for a fixed 5V supply you would want the LM2596-5.0. In the UK Farnell (element14) list them: see here, but I'm sure that there are many other suppliers.
The fixed output voltage version using the LM2596-5.0 requires only four other components: input capacitor, "catch" diode (eg 1N5821, 1N5824 or suitable SMD), inductor and output capacitor.
There is also a lower current DC-DC switcher: the LM2575-5.0. Farnell listing here. Again it requires only four other components.

See the data sheets for full details of both devices. Talking of which, I used the On Semiconductor and Texas Instruments data sheets.

Mark

[Dave]

http://www.ti.com.cn/cn/lit/ds/symlink/lm1575.pdf page 10

Given a 5 volt 200mA draw, at best the LM1575 will be about 78% efficient.

This is the device I selected for the other cards in the range:

http://www.aosmd.com/res/data_sheets/AOZ1016AI.pdf

At 200mA it starts at 85% efficiency. I have kitted these already for other cards, but I squirm at the idea of putting these on a very small memory card that will usually draw current below threshold and be very RF noisy. Ideally, I need a buck regulator rated around 600mA.

The problem is I have 500 78M05 in DPAK, fully kitted with caps, and they're essentially free to use as they'll never be applicable anywhere else.

How about this? I think about 50-75% of SuperRAM sold will replace a 512K DRAM memory expansion. The savings alone from replacing 16xDRAMs with 2xSRAMs will more than make up for the 15% of inefficiency of the voltage regulator over more modern methods. The new 7805 will still be 40-50% more efficient than an 80s 7805.

For some perspective, based on 5v output, 200mA, 12v input:

Code: Select all

0%..:..10..:..20..:..30..:..40..:..50..:..60..:..70..:..80..:..90..:..100%
                        1----1                    2    3    4

     1. 80s 7805    2. modern 7805    3. LM1575    4. AOZ1016AI

[1024MAK]

Dave, I have never had a problem with you using a 78M05.

I also don't have a problem with switching regulators being used. The examples I gave earlier were just that - examples. There are plenty of other regulators of various types.

The only point on which I disagree with you is the way that various different "linear" regulator systems are described as being significantly different in their efficiency.

You have your option. I have mine.

What, you still reading - go and design the layout.... I'm waiting...

Mark

[Dave]

*grins*

Thanks, Mark. I know, there's many ways to skin a cat, amiright?

The design is coming along smoothly. Everything is on schedule.

I also sent a private message of encouragement to tcat. They're happily doing their own thing as a learning exercise. I think that's great, and should be encouraged. I look forward to seeing their evolving design here. Hopefully, people won't be so forceful in their opinions about tcat's design decisions. Hobbyists should be encouraged.

[tcat]

Hi All,

Do you mind following this link, I feel this went slightly off topic, I wish to make a "leap" from GERT INTERNAL design to DIY board design and recommended components list.

http://theqlforum.com/viewtopic.php?f=2&t=877&start=10

Many thanks
TCAT