Discussion: Anything wrong with the MyMensch SBC?

[tokafondo]

It seems to be an updated, upgraded, enhanced w65c256s. It even has one of the features I miss from the '265: being able to program it with custom boot code instead of relying on an external ROM.

It also has enough internal RAM to not to consider it a toy and different interfaces that allows to communicate with devices so the address and data bus can be left untouched.

Is anything wrong with this design?

Thanks for your replies.

[GARTHWILSON]

I've been meaning to give it a try when I get some time.

[rpiguy2]

tokafondo wrote:

It seems to be an updated, upgraded, enhanced w65c256s. It even has one of the features I miss from the '265: being able to program it with custom boot code instead of relying on an external ROM.

It also has enough internal RAM to not to consider it a toy and different interfaces that allows to communicate with devices so the address and data bus can be left untouched.

Is anything wrong with this design?

Thanks for your replies.

I think the main problem is that it is a lot of money for a $5 Intel FPGA on a $1 PCB.

It is also odd that the internal clock speed is constrained to 14mhz. If all the RAM is on the FPGA there is no reason to restrict the clock or not make it selectable.

[kakemoms]

Were to begin...

1) Its too expensive. Put out a base price that is comparable to the component cost and license the software separately. Most SBC's have alot on the board, but the Mymensch is more like a PicoPI or Arduino nano at the moment.

2) Power requirement. The 65x02 series is known for its low current consumption. How does this solution compare? I can't find anything about current in Rev-C. Its superimportant.

3) Make the oscillator frequency source a choice for the user. Either internal oscillator or external. 14MHz is comparable to very cheap arduinos. You can get a STM32 card that runs >100MHz. A Risc-V runs at several hundred. If you don't have 2), you need this.

4) 5V interfacing. The 65x02 series can go there, so why change to this at same speed?

5) Its often about shields. The Arduino has them, The Raspberry PI has them, most STM32 SBC's has them. Today we want to put on a display and get something up and running. All schools here use Arduino if speed isn't that important, and its because you can plug in everything from gas sensors to displays and lasers.

6) Software. Make it Arduino IDE compatible to get a chunk of that market. Most other SBC's can be controlled from that, either through a supplier's library or a third-party library. Its always free and its relatively easy to use.

7) More software. Think Lego mindstorms. They are super-easy to program for anyone. Click-n-drag objects to get functionality. Graphic IDE.

If you want into R&D market you need 7), but also interfacing. Think Labview. Researchers know almost nothing about programming these days and need solutions that work. Sensor data and stuff like that. Not everything needs 10GS/s and a i9.

[tokafondo]

Ok. What about the 6502 part of it?

[BigEd]

Just to recap, the MyMensch SBC offering is a small FPGA-based 6502 dev board, 3.3V I/O, 14.7MHz clock, for $65. There's a corresponding mainboard to go under it, the MyMENSCH PRO Board, for $28.

One might view this as an educational offering, to be bought in the dozens by colleges or similar. Or as a dev board for someone planning a design which would use WDC's IP. Or as a hobbyist board, which is probably where most of the audience here are coming from.

It might be that the board is well-designed and easy to work with. I think it would need to be, given the availability of open source 6502 cores which run at 100MHz on cheap FPGA dev boards. I think that's what I'd compare it with. It might be that the $65+$28 is quite a good price for the amount of I/O you get. But most likely, competing solely on price isn't the thing, because it's the ease of getting going and unlocking the potential that's important.

It's possible that the SBC being FPGA-based isn't important - maybe you're not expected to customise the on-chip hardware, only to use it as-configured. In which case it would be good to know what that hardware looks like.

I see there's a getting-started guide here:
https://wdc65xx.com/gettingstarted/myme ... g-started/

If that doesn't answer the questions, that's not a good sign.

(It's possible the 6502 core could clock at much more 14MHz, using an on-chip PLL. It's not made clear.)

[tokafondo]

So, we have a compact SBC that features all of this:

Several standard, "classic" 65xx I/O ports and several extra stuff like SRAM, flashable ROM, ADC, hardware multiply and divide units, SPI, timers and a bunch of I/O lines, already mapped in the first 64K of RAM.

And everything handled the 65xx way, so any existing code could be recycled for them, and the standard tricks that hardcore programmers use could be applied.

Isn't that a great thing? I mean... It seems most people complain about the use of cheap hardware, sold at premium price but... Would be all of that easy to implement in a blank FPGA anyone could buy for a couple of coins?

[BigEd]

That does sound like a good mix of micro-controller type peripherals. (I'd be particularly interested to know how it presents the multiplier.)

So, yes, it's all there for you, and no messing with FGPA toolchains. And the price is comparable or better than the least expensive FPGA dev boards. (The price of the FPGA chip is a minor part of that.)

On the other hand, anyone who enjoys creating and building FPGA designs can build their own, should they wish.

[GARTHWILSON]

Shucks, neither of WDC's websites is coming up for me right now to see how they did the A/D's; but 9 inputs of 12-bit A/D converter probably puts it in a category above what most FPGAs will do.

[John West]

GARTHWILSON wrote:

Shucks, neither of WDC's websites is coming up for me right now to see how they did the A/D's; but 9 inputs of 12-bit A/D converter probably puts it in a category above what most FPGAs will do.

Ten years ago, yes. But almost all of Xilinx's current generation of FPGAs (the exceptions are the two smallest Spartan 7s) have 12 bit ADCs with a decent number of multiplexed inputs built in. I assume Intel's offerings have something similar.

[rpiguy2]

GARTHWILSON wrote:

Shucks, neither of WDC's websites is coming up for me right now to see how they did the A/D's; but 9 inputs of 12-bit A/D converter probably puts it in a category above what most FPGAs will do.

ADC appears at $7FB0 and is on pins J4-9, 11, 13, 14, 15, 16, 18, and 20.

How to access it at the most basic level is described on page 28 here:

https://www.westerndesigncenter.com/wdc ... 1M16SA.pdf

[Jmstein7]

BigEd wrote:

That does sound like a good mix of micro-controller type peripherals. (I'd be particularly interested to know how it presents the multiplier.)

So, yes, it's all there for you, and no messing with FGPA toolchains. And the price is comparable or better than the least expensive FPGA dev boards. (The price of the FPGA chip is a minor part of that.)

On the other hand, anyone who enjoys creating and building FPGA designs can build their own, should they wish.

Exactly. And, if you think about it, it really ends up being a good deal. Consider... from Mouser, a 65c02 will set you back $10.75, a VIA is another $10.75, an ACIA is $7.27, and about $3.36 for a decent 32k of SRAM (the rev-c has 42k). The rev-c has 3 VIAs and 3 ACIAs - that's $32.25 and $21.81. And, you can easily find one 32k eeprom for about $2. So, all in all, in discrete parts, that would be just over $70US. And the rev-c has so much more than that. I have one, along with the breakout board, and the "soft" logic it uses works just like its "hard" counterparts.

As we all know, it is very, very hard to find free "soft" cores that a) work together "out of the box" and b) work as if they were "drop in" versions of the same. To get something like that working takes a lot of effort and a reasonable amount of knowledge. WDC gives that to you out of the box.

And, of course, this is not really for use as an arduino replacement - it is a dedicated 65cxx machine with contemporary extensions.

Jon

[plasmo]

Don't forget the cost of pc board, assembly, defective boards, marketing, and supports. A simple rule of thumb for the price of a product is multiply cost of its parts by 2.5.
Bill

[Jmstein7]

plasmo wrote:

Don't forget the cost of pc board, assembly, defective boards, marketing, and supports. A simple rule of thumb for the price of a product is multiply cost of its parts by 2.5.
Bill

Precisely. And, of course, it doesn’t hurt that it is well-made, either. And, WDC is also very responsive to customer inquiries.

Just my experience with them.

Jon

[drogon]

plasmo wrote:

Don't forget the cost of pc board, assembly, defective boards, marketing, and supports. A simple rule of thumb for the price of a product is multiply cost of its parts by 2.5.
Bill

Lets not also forget the cost of shipping, import duty and local tax. I can build a 65xx system from scratch in the UK for less than it would cost me to import one of these.

I can't find any UK distys - anyone?

Also, I can't find the cost of shipping from the US unless I more or less fill in the first 2 pages of their order form - and even then I know I'll be hit for 22.5% combined customs duty and VAT on top of postage and item cost.

MyMensch for > £100, Or an Arduino/Pi Zero for < £10, or a Pi Pico for < £5 on a board, chip for £1. Not exactly a hard choice to make.

So while I'd like to get my hands on one of these it's simply not going to happen as it's just far too expensive )-:

-Gordon