Looking for a "reliable" source of EPM7064SLC44

[Dave White]

I'm starting on a new project based around the 6809 and planned to use an  EPM7064SLC44 for general glue logic. I'd bought ten off of e-bay a year or so ago and all but one were jtag locked and couldn't be unlocked using the 12V trick.

I'm looking for a few good chips, but don't want to risk e-bay again. Does anybody have any experience of a reliable source for these chips, or even the EPM7128? Preferably in the US, but as I'll be visiting family in the UK in August, a UK source would be fine too.

Thanks

[Bill Shen]

UTSource.net has EPM7064 and 7128 that are good even used parts.  I’d advice you have a working CPLD programming setup ready to check out parts so you can return them if parts are not programmable.

  Bill

[Bill Shen]

I’m interested in your 6809 project.  I’m just started working with 6809.  I’m not familiar with it so it is a steep learning process.  I do have a working 6809 SBC under development which has EPM7128 as the glue logic.  https://www.retrobrewcomputers.org/forum/index.php?t=msg&th=700&goto=10010&#msg_10010
  Bill

On Sunday, July 17, 2022 at 12:36:59 PM UTC-6 Dave White wrote:

[Dave White]

Thanks Bill. I've just ordered 15 pieces from UTSource, so hopefully these will be fine.

My 6809 design is in it's early days right now. I have a bunch of 6808E and 6309E chips in PLCC packaging but had a heck of a time finding the pinout for the PLCC "E" versions. But I've finally managed to track down the information I need, so I'm busy researching designs. I've always been a Z80 man myself, so this chip is brand new to me too, but it's always been on my wishlist. When I get a bit further along, I'll post some information up to my web site.

[Alan Cox]

On Sun, 17 Jul 2022 at 20:52, Dave White <mrgcm...@gmail.com> wrote:

Thanks Bill. I've just ordered 15 pieces from UTSource, so hopefully these will be fine.

My 6809 design is in it's early days right now. I have a bunch of 6808E and 6309E chips in PLCC packaging but had a heck of a time finding the pinout for the PLCC "E" versions. But I've finally managed to track down the information I need, so I'm busy researching designs. I've always been a Z80 man myself, so this chip is brand new to me too, but it's always been on my wishlist. When I get a bit further along, I'll post some information up to my web site.

I did a board for the 6809/6309 standard version with the 40pin package for RC2014, in that form it's really simple but I too couldn't find any decent info on generating the clocks or the PLCC pinout which is annoying as I've got a C09E and a couple of other parts that would make a nice dual CPU system.

It's a nice chip and you can run some interesting things on it - Flex9, NitrOS9, Cubix etc.

 

[Dave White]

It actually turned out to be straightforward, I'll send you the info when I'm back at my computer.

[Bill Shen]

Alan,
What hardware are expected to run Flex9, NitrOS9, or Cubix?  I assume they need a mass storage device, serial port, timer, possibly banked RAM?  I'm looking into porting Assist09 monitor which needs a serial port and timer/counter (6840).  Easy enough for me to build a expansion board with 6840 and CF disk, but I like to add whatever hardware needed by the operating system. 

  Bill

[Alan Cox]

On Sun, 17 Jul 2022 at 23:39, Bill Shen <coinst...@gmail.com> wrote:

Alan,
What hardware are expected to run Flex9, NitrOS9, or Cubix?  I assume they need a mass storage device, serial port, timer, possibly banked RAM?  I'm looking into porting Assist09 monitor which needs a serial port and timer/counter (6840).  Easy enough for me to build a expansion board with 6840 and CF disk, but I like to add whatever hardware needed by the operating system.

Cubix: 8K or so ROM, serial or video console, a storage device and some RAM. No timers, no interrupts, no banked RAM. It's also logical block based so the BIOS is really trivial for a CF adapter - see my 6809 ROM as it's what my 6809 boards boot into if they don't find a disk boot block to use. It's got echoes of VMS and the like and has a single layer of subdirectories. It's all 6809 assembler and can even self host and includes a subset C compiler as well. Open Source. It (and Flex) can be made to work with the usual RC2014 banked memory card by copying the "ROM" into the top of RAM and going all RAM with just an I/O window generated by the CPU card.

Flex: even less but does need RAM in certain address ranges (look up 6809 "Flex Adaptation Guide"). Flex occupies about the same historic space as CP/M did but in the 6800 and 6809 world and with a very different mindset and way of doing things sometimes. The 6809 one as supplied was about CP/M level in functionality although with a better disk format. Third party addons improved it and even added hierarchical directories. Ownership not entirely clear.

NitroOS9 for the full level wants timer interrupts, rtc, banked memory, and preferably video as well as storage. It's derived from OS/9 which is a sort of marriage between a heavily cut down unix and an rtos. It's very very neat and in its day OS/9 was seen as a really serious system. OS/9 68K fared less well and it never really went anywhere after 68K died. NitrOS9 however has a big active 6809 community who mostly run it on Tandy COCO 3 and the like. It's also all open source. I've not yet ported it (too many other toys) but did build a 68/6521 and 6840 card for the job and for Fuzix.

If you are working in programmable logic also take a look at 65SPI, it's a high speed SPI interface for motorola bus systems and rather than all the legacy CF stuff can drive an SD card at system memory speed. It's used in stuff like the Dragon NX32 and Mooh, as well as a bunch of 6502 SD interfaces.

[Dave White]

Okay Alan, I've posted the 6809E information to my blog - http://spectrachrom.com/blog/6309e-6809e/

[Dave White]

And it seems that Google may be mangling that address, so if it doesn't work, just copy the link address and paste it into a browser until I can figure out what Google is doing.

http://spectrachrom.com/blog/6309e-6809e/

[Sergey Kiselev]

You can use Atmel ATF1504AS instead. Atmel provides the POF2JED utility to convert files from Altera format to Atmel.

It is likely out of stock right now though, due to chip shortages, but at least it is still manufactured and supported...

[Dave White]

Thanks Sergey. I was aware of that option, but I wanted to bypass the conversion step and program directly from Quartus. And you're right, when I looked for those chips at the "big name" suppliers, they had 0 stock.

[Dave White]

Well that didn't go as planned. UTSource e-mailed to say the supplier I ordered from was sold out (then why list them as available?), but they found the parts at other suppliers at almost 3x the price - I don't think so. Now we'll see if they refund the payment.

[Bill Shen]

Are in located in USA?  I can sell you up to 10 epm7064slc44-10 from my stock for $3 each.  Shipping in USA is $5.  I’ll test them to make sure it’s jtag is accessible.

  Bill

[Dave White]

I'm in Minnesota - yes, that would be great! 

I still have your Paypal info from a couple of years ago, but if you PM me with the address to confirm, I'll send the payment and my address.

Thanks!

[Richard Deane]

I had the same problem with UT source, product described as available but out of stock and cancelled as soon as I placed the order.  Strangely , but good, that I have never had that problem with AliExpress, from whom I have placed about 20 different mixed orders for retro components over the past couple of years - for cpu, memory, crystals, sockets, chokes etc and not had any problems of the parts not working.

Richard

--
You received this message because you are subscribed to the Google Groups "retro-comp" group.
To unsubscribe from this group and stop receiving emails from it, send an email to retro-comp+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/dedf0003-b078-4e2b-bff2-52f312804b97n%40googlegroups.com.

[Bill Shen]

Early this year I was ordering multiple items from UTSource and also encountered the cancellation of one item for the first time.  The alternate was about twice as expensive.  I did ordered the alternate but at reduced quantity.  Items all arrived and are OK, but the last-minute cancellation was somewhat disturbing.
  Bill

[Mark T]

I have also had the same on multiple occasions. I was suspicious that Utsource were buying the stock from the cheaper source and adding to their own stock at the higher price.

[Pellatonian]

Back in May I placed an order for 30 pcs AT27T512R EPROMS at $3.01 per piece from UTSource, which was accepted.

Got this back 1 week later:

"Thanks for your payment, but we are sorry to let you know regarding order item AT27C512R-45PU ,

our supplier just called us it was nearly sold out, we tried from other suppliers,

but the price is $100   /pcs, Can you accept it ?  If yes, pls pay us below :

 100*  30-90.2634=2909.7366  

Pls pay $2909.74    to our paypal account ......"

             ^^^^^^^^^^^   !!!!!

[craig a]

UTSource is just a chip broker
They have minimal inventory, they mostly just maintain lists and inventory counts for others and do commission sales for suppliers that don’t want to bother with small scale sales.  If their supplier’s inventory or published prices are off then UTSource has to adjust.  

[Dave White]

Apart from the device and the price, that's word for word what they sent me - the price more than tripled. Sounds a lot like bait and switch.

Did you have any problem getting a refund?

"Thanks for your payment, but we are sorry to let you know regarding order item EPM7064SLC44-10 ,

our supplier just called us it was nearly sold out, we tried from other suppliers,

but the price is $ 7.5  /pcs, Can you accept it ?  If yes, pls pay us below :

     $7.5*15pcs-$34.47=$78.03

Pls pay $78.03  to our paypal account ......"

[Sergey Kiselev]

UTSource is full of crap. They list parts in large quantities "in stock", for low/OK price, and once you submit an order, they increase the price on you.

I've stopped using them for this reason.

Also, I'd like to reiterate the point, that by buying the "recycled"/"remarked" parts, that are currently available from official manufacturers, you are supporting Chinese chip counterfeiting industry, and close-to-slave labor...

If you can afford, pay a bit more, and order parts directly from the manufacturer or the official distributor... 

In addition, convert your design to use currently in production parts - e.g. Flash ROMs instead of UV EPROMs...

On Sunday, July 24, 2022 at 7:09:52 AM UTC-7 Pellatonian wrote:

[Bill Shen]

The recent trend with UTSource is certainly alarming and discouraging.  I don’t think I will recommend UTSource anymore.  

One should bear in mind that “retro computing” involved obsoleted parts, many are no longer in production;  they are effectively used parts without chain of custody so every build has the potential for hidden defects even build with parts removed from working hardware.  Of course relabeled parts add yet another level of hidden defects to successful construction of retro computers.  The situation will only get worse with every passing year as limited parts are used up.  We have enjoyed the renaissance of retro computers with the abundance of recovered used parts mostly from China, but the supply chains are becoming unreliable.  Experienced users can still acquire likely working parts through network of known good suppliers and incoming inspection and testing, but I afraid it will be increasingly difficult for new users to enjoy the retro computing experiences.

  Bill

[Richard Deane]

Due to the parts issue, I have redefined my personal view of what retro is. I now consider programmed items like CPLD, FPGA or standard components in smd form,  as OK to implement the logic of genuine retro parts, and perhaps to save space, but if the component actually runs software at run time to emulate something then it is no longer proper retro. Having said that I enjoy and see a legitimate space for products like altairduino, runcpm, rc2040 (rc2014 emulation on rp2040), esp32 with fabgl altair8800, and hybrid products like Z80-MBC2.

I am reminded that I did have difficulty claiming back from UT Source when I tried to cancel my order. I am sure they were price gouging. I would never, ever, use UTSource again. My experience with AliExpress has been good. 

Here in the UK, post Brexit, I find that it is not feasible to purchase small hobby quantities of components from local suppliers and I have to use China. Ebay UK sells what comes from China anyway. I batch up to order from Mouser from time to time, but they don't have everything (e.g. Resistor networks as needed for etched pixels serial card)

To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/95046fa1-0c57-43cf-bef2-d8d851e492c0n%40googlegroups.com.

[Alan Cox]

> sells what comes from China anyway. I batch up to order from Mouser from time to time, but they don't have everything (e.g. Resistor networks as needed for etched pixels serial card)

You should just be able to use a line of 0.125W vertical resistors
instead for the ones protecting the signal lines.

Alan

[Richard Deane]

Thanks for the tip! 

So far I have compromised with a shorter resistor pack for RN1 and RN3 with the last two holes exposed (not essential pins for my needs, if I remember correctly), and had considered that I could add more individual resistors later if necessary.

It was just annoying that each supplier seemed to have a different subset of RN components, even Aliexpress being quite poor, and not always clear about the terminology for the type of RN.

The RNs are tidier.

Cheers

Richard

[Chris Swan]

> Here in the UK, post Brexit, I find that it is not feasible to purchase small hobby quantities of components from local suppliers and I have to use China. Ebay UK sells what comes from China anyway. I batch up to order from Mouser from time to time, but they don't have everything (e.g. Resistor networks as needed for etched pixels serial card)

I've found Bitsbox (https://www.bitsbox.co.uk/) pretty good for getting components in the UK. Their catalogue isn't super comprehensive, but they have most of the things I need, most of the time (and they do have resistor networks, though I can't be sure whether they're the right one's for Alan's card).

[Richard Deane]

Thanks, I will give it a look,

Richard

To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/64e74af4-cafa-43a7-8883-b76088d1a27an%40googlegroups.com.

[Earth Person]

Hello Dave,

Went the eBay route, and went UTSource a month ago as it I've had good results. Was bit with the comments later where they have one price then they come back. So I went with the used parts and went on the edge and bought 100 of the EPM7128SLC84-10N's. At 100 Qty I was quoted 4.50 USD. So far they all work. Maybe I hit a goldmine. Doubt it. We'll see. I use a lot of these between the Altera, UP1 and UP2 boards, along with the CPLD trainer. I'll add these to retrobrew inventory. It's a dog eat dog world, so don't be a dog!

I did put in the order about 2 or 3 weeks ago, and I do big orders, so it did take some time to find the quantity I needed. New? No idea. It was stated used. Assume the risk.

Also found out, the eBay route also went to UTsource. So, extra long time. 

Sincerely,

William

[Dave White]

Excellent - I'll keep an eye on the inventory page. It would be nice to play with some bigger CPLDs with more I/O.

[Bill Shen]

Since CPLD replaces so many TTL logic and some serial, timer, or general IO, it may worth the efforts to establish known good sources as well as a stockpile of part inventory in USA and Europe where shipping can be reasonable and fast.

  Bill

[Earth Person]

Greetings,

I have updated the wiki with my contact info and inventory status of the EPM7128SLC84-10N's.

Sincerely,

William Harrington

[Thomas Winkler]

I'm also working on a 6809/6309 SBC.

And I'm using also this kind of CPLD.

But currently I'm not using the EPM Typ but this compatible ATF CPLD from Atmel.

Both CPLD (EPM7xxx and ATF15xx) are available at AliExpress.

I love this CPLD and also the old GAL, cause it is simple to program and it can work with 5V Logic without level shifter.

The 12V unlock is working fine on my ATF1504.

I'm sure it works also on this compatible EPM types 

---

My SBC exists in hardware and I also have made an Emulator for it.

Currently it runs CUBIX 1.3.
But my wish for the future is running OS9 or NitrOS9.

My SBC v1 has an MMU 6829 which is able to address 2MB of memory.

Cause it is difficult to get this Motorola MMU I have made my own MMU: MMU-16

MMU-16 is able to address 16MB of memory.

It is fully compatible with MMU 6829.

MMU-16 hardware ist based on:

- 2 very fast 16K SRAM (chache RAM) for memory mapping

- 2 CPLD  (ATF-1504-44 and ATF1504-84)

MMU-16 can do same as 6829 but:

- it supports 16MB memory  (but 2MB on 6829)

- it supports 256 tasks  (but 4 tasks on 6829)

- page size is 2KB like on MMU 6829

best regards

[Bill Shen]

What tool you used to design ATF150x?  I’m only aware of WinCUPL and translations of Quartus design to ATF using POF2JED.  Both approaches have problems so I’m looking for a better solution.

  Bill

[Bill Shen]

Lack of good tool for ATF150x is ultimately why I continue to work with Altera EPM7xxx with Quartus tool.  With patience and lucks, EPM7xxx can be acquired in quantity significantly cheaper than ATF150x which are getting more expensive and harder to find currently.  CPLD is capable of complex state machines and buried ROM which Quartus can handle with ease but very hard to do with WinCUPL.  I have acquired enough EPM7xxx to last my lifetime.

  Bill

[Tom Storey]

The only other tool I know of for the Atmel parts is something called ProChip Designer. I believe it can handle Verilog and VHDL (probably not System Verilog, though), but I believe the simulation tools are not included or free to obtain, so that pretty much wrote that off for me as a solution. Simulation is a HUGE part of programmable logic development for me.

So I stick with WinCUPL and Quartus II as both of these have freely available simulation tools, albeit both with their own quirks. Ive also built up a bit of a collection of compatible Atmel and Altera parts, so I am leaning more towards Verilog these days, and lesser towards CUPL except for GALs.

But Im also starting to play around a bit with some Xilinx Spartan II FPGAs. These are 3.3V devices, but when the IOs are configured for the LVTTL standard they are 5V tolerant. They will of course only output a maximum of 3.3V but if you look closer you'll find that a lot of 5V CPU, memory, and other peripheral chips will accept down to around 2V for a guaranteed logic high. This is probably so that they are compatible with TTL families like LS etc. So 3.3V output is perfectly fine to drive these guys and where they arent you can buffer the signals up to 5V using some HCT logic for example.

I have a small stock of XC2S200 Spartan II FPGAs which have a few hundred thousand "gates", so theres quite a lot of functionality I'd be able to squeeze into them for some bigger more complicated designs, with a little bit of buffering here and there perhaps. The biggest challenge with the ones I have are that they are BGAs, but my reflow soldering technique is getting better, and with a hot plate I think its going to be achievable. First step will be to see if I can make a breakout board for one, and Im looking at implementing the configuration PROM in a PIC microcontroller with a SPI flash memory attached.

The XC9500XL series of CPLDs are also 3.3V IO but with 5V tolerant IO, and I believe are EE devices that can be programmed via JTAG and dont need any configuration PROM to get going.

So, for me, the prospect of using 3.3V logic devices in my 5V circuits is looking more and more real going forward. It will take a bit more effort to design with, but the idea is opening doors.

Tom

[Tom Storey]

Oh, before I forget, there is an effort to reverse engineer the ATF150x series to make them compatible with open source toolchains like nmigen, which would then allow synthesis and simulation/formal design using a variety of other tools.

It looks like they have completed the 32 macrocell device, with the 64 and 128 still WIP.

[Sergey Kiselev]

I am wondering that are the issues (I think Bill mentioned something?!) with using Atmel's POF2JED tool to convert the Altera/Quartus II designs to ATF15xx?
I've used that for a few projects, and it worked well. One time it didn't work (IIRC, due to a pilot error), I've contacted Atmel (p...@microchip.com), and got a bit of help from their application engineer.

[Tom Storey]

Ive also not had any issues. There are quirks with both of them, but for the most part I can just carry on and build stuff.

Personally, Im not really sure there is anything better out there than the Quartus II + POF2JED solution.

--
You received this message because you are subscribed to the Google Groups "retro-comp" group.
To unsubscribe from this group and stop receiving emails from it, send an email to retro-comp+...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/d0e45e44-c03f-4ee6-8a63-7416a5b87aben%40googlegroups.com.

[Bill Shen]

POF2JED translation seems to work well for low-to-mid utilization of CPLD logic; I've translated a whole series of EPM7128S designs to ATF1508 successfully, but these designs were not nearing 100% logic utilization.  Simple designs for EPM7064S also translated OK.  However, I have several very dense designs (ZRCC, CRC65, VGA6448) that utilized logic beyond 100% and required optimization to reduce it back to 100%; they didn't translate to ATF1504 successfully.  My unproven observation is EPM7128S -> ATF1508 seems more likely to succeed than EPM7064S -> ATF1504.

  Bill

[Tom Storey]

What sort of utilisation were you translating?

In recent times I designed a VGA CRTC which used 121 out of 128 macrocells (target was a EPM7128/ATF1508) which is 95ish% utilisation, and other than some issues with fitting (maybe not helped by me wanting certain pins in certain locations on the chip) I didnt have any issues translating the design with pof2jed. The design worked fine on both target chips.

To view this discussion on the web visit https://groups.google.com/d/msgid/retro-comp/6e97c263-b0e3-47fa-ad5f-88a5c63c03a4n%40googlegroups.com.

[Bill Shen]

My experience with ATF150x is limited 84-pin ATF1508 and 44-pin ATF1504.  I don't have too many designs with 84-pin 128 macrocell ATF1508/EPM7128, but every translation attempt was successful.  The most complex one has 100% pin utilization and 124 macrocells used out of 128 macrocells.
http://forum.6502.org/viewtopic.php?f=10&t=6974&hilit=cpld+trainer&start=15#p91124

I have quite a bit more designs in 44-pin EPM7064 but have also encountered many problems translating to ATF1504 except for very simple designs.  I do try to pack as many functions as possible so design like ZRCC is 100% utilized in logic as well as pin and VGA64RC is 63/64 in macrocell utilization and 100% pin utilization.

Thinking about it now, I wonder whether it is related to package type; I should try to squeeze more functions into EPM7128 and see if that all translated correctly to ATF1508; conversely I can try remove functions from EPM7064 that didn't translate and see what needs to be removed for the translation to work.
  Bill