Email: Password: Remember Me | Create Account (Free)

Back to Subject List

Old thread has been locked -- no new posts accepted in this thread
Richard Erlacher
11/13/11 01:09
Read: 604 times
Denver, Co

#184709 - I think you're confusing two different parts.
Responding to: Per Westermark's previous message
Per Westermark said:
But note that the original EPROM chips did not support electrical erase. And applying UV light meant bulk erase.

The P8052 didn't have any EPROM. There was a P8652 that was OTP, but, being in the plastic package, it couldn't be erased in any sense.

So the use of a single external EPROM would not allow the lower part of that EPROM to contain a loader handling any serial port and then have that loader receive more program to burn into upper parts of the EPROM - especially since the EPROM would not run the processor while being programmed.

Yes, I have to agree, since that's what I previously said.

Even most flash chips have the limitation that they can't serve out code while having other flash sectors being erased/reprogrammed.

Internal flash or otherwise reprogrammable program store is not a factor here. External flash is probably not a good choice either, as it requires intelligence to program it. EPROM is dumber, hence easier to program, particularly with small parts. FRAM is easier yet, the memories can be quite large. Though they're not as large, BBRAM's are pretty easy, too. The fact is, we don't know what hardware the O/P has at his disposal. He could have equipment available in a lab at school, or with a friend, or whatever.

So the simplest solution would be to have two memory chips. One EPROM or flash chip with a boot loader. But how (?) would that program get a boot loader when there are no stand-alone programmer available to program it?

I agree. That's why I mentioned the BBRAM or FRAM. Those require no programmer. If the nEA line were then tied low, the chip would attempt to use only external program store, enabled by nPSEN, and, with proper hardware design, (not rocket science, as it requires only that nRD, nWR, and nPS be negative NOR'd to produce a negative-going strobe at the occurrence of any one of the three, by which the external memory data buffer, if there is one, is enabled, and the external memory as well.) What this all amounts to is degrading the operating mode of his MCU to that of a P8032, which uses only external memory. If, then, he has the means to program an external EPROM or even just to write to an FRAM, which is all that's necessary and NOT rocket science, he could locate the bootloader there, serially load the code he wants to use in the FRAM, and then move it to the appropriate region in memory so that, next time he boots the MCU, the code he bootloaded will be located properly. FRAM's are available in large size, BTW. I've used 512 kB sized ones with wire-wrap adapters, to provide a big enough nonvolatile memory. By creative use of external logic, i.e. a few gates and jumpers, he could manipulate the high-order addresses to make the bootloader appear in the correct place for bootloading, yet move it out of the way to run his newly-loaded program. Clearly there are many options. No matter what he does, though, he'll not be able to erase or reprogram this particular part.

The second memory could be a RAM or a flash, and could get the contents by the boot loader program.

But no - there would not be anything simple about the above. So we would be back to my original post.

When you don't know how to do something, you always find it difficult. Once you know how, it's easy. That's especially true with digital hardware.

The cheapest way to handle this would be to beg RAMTRON for a sample of one of those FRAM's and an adapter for the TSOP-II to a more convenient footprint. Then, with a parallel port on his PC, he can build a device with which to write to it. Even easier, he could use an EPROM, the easiest of which is a 2716, of which there are still a few around. A programmer for one of those would consist of a few counters (3 4-bit counters, perhaps) and a one-shot to time the maximal 50 ms time requirement for it. Then the parallel port can, byte-by-byte, be used to program the EPROM from the parallel port, using one input bit to monitor the post-write timeout provided by the one-shot. The whole thing would cost less than $10, even at today's inflated prices.

Yes - a programmer can be written. Yes, the processor can be used. But it would cost. It would take time. And it would probably not work because it would require too many individual steps to be correct. And someone who starts from scratch would not have the equipment and experience to figure out all the broken steps until the design finally started to work.

You're right, it's not for the faint of heart, nor is it a task for a total newbie, as he indicates he is, but it's not a budget buster.

A fraction of that time and a fraction of the cost, could be used to buy a development board with a modern processor with a RS-232 or USB interface where a PC program could send a new program. And with a modern processor that would have internal, hardware, support for debugging of the downloaded program.

Well, that's probably true. If he's bent on using THIS PARTICULAR DEVICE, he's got a hard row to hoe. If, however, he has a the intelligence to manage the task in the first place, he'll quickly see that he needs, at least, to obtain a serially programmable MCU, costing less than $2 or so on eBay
for example, and be done with it. It's not certain that he even requires an evaluation board, provided he has done the research to enable him to buile the programming circuit, usually consisting of little more than a MAX-3232A and a serial connector.

Going the new route, would mean lots of time available to learn new skills on a working platform.
Going the old route would give lots of experiences that may not be wanted at this time. Like 100 different things that can go wrong ending a project in failure.

That's what I'd use in this sort of situation.

And what situation would that be? You starting from scratch, without any existing software? No existing hardware but a dumb processor and one or more UV-erasable EPROM chips but without any UV eraser? No tools to fill that EPROM with contents? No tools to drive raw processor pins to use as a building block for a programmer? No tool capable of receiving RS-232 data from a PC to send into latches to then (together with pulsing of address counters) program the EPROM?

That's the situation that confronted those of us who wanted to use the 805x's back in the early '80's. There were tools, of course, but they were VERY costly and cumbersome, There were no IBM PC's yet ... at least not evolved enough to be of use, and the tools Intel provided cost more than a new Benz. We had to read the datasheets, write our own software, and build our own hardware. Fortunately, the datasheets included programmer schematics for the EPROM-based parts. We hadn't learned about microprocessors or microcontrollers in college because they didn't exist back in the '60's. We had to do exactly what this fellow, the original poster, has before him. I'm of the opinion that we must credit him with at least as much intelligence as I had back then. Perhaps not as much as you possess, as you always seem to know better, but a reasonable amount. As a result, I have to believe he's at least clever enough to do what I did back when it was totally necessary. Back then, there was no 8052.COM.

Anyone who gets to this forum with the question the OP had, would be in a position where it doesn't matter what "can" be done. The important thing is that with lacking tools and lacking experience, lots of things that can be done just shouldn't be attempted. Becaus the cost would be too high. Because the failure probability would be too high. Just because other - very cheap - alternatives would allow much more learning in way shorter time.

If you have a $300 old car, it isn't meaningful to invest $3000 to try to get it to run a bit better.

If you have the skills and a truly unique wreck, it might be meaningful to invest $30,000 instead. Just because you will be able to create a museum-class vehicle in mint condition.

This is not a thread where an experienced developer is trying to recreate an important historical artifact, but someone who have a "wreck" on his hands and isn't realizing that the cost it would take to make it usable is higher than the cost of buying a new SDK complete with documentation.

You're not the first to suggest that this is a fool's errand for one who's totally inexperienced. However, aside from my own belief that the O/P needs to consider his goal before deciding what to do, he should also consider that he can build the same circuit with a modern, serially reprogrammable, device, for which PC-based serial loaders are available from the chip manufacturer, that could be built with the old Intel part, he may actually prefer to suffer the consequences of starting from scratch. Until he tells us what he wants and why he wants it, we can only guess.

I believe he should be encouraged to do what he wants, and not be dissuaded because there's an easier way. People have walked up the half-dome in Yosemite national Park. Others have climbed the face. It's just a personal choice.

As for your car analogy, I routinely run across situations where someone has an old car worth less than $300, and wishes to spend two years and $30k on it to restore it to "original" condition. Sometimes that's because it is of the same vintage as they are, and in others it's because it was the first car they learned to drive. I've seen TV programs on which they spend $100k on a car worth $50 to restore it, even though the restored value is on the order of $10k. I've seen TV programs on which they take a new, $50k vehicle, modify it at an expense of $200k, to make it into a vehicle worth no more than $10k. I can't make any sense of it. Can you?


List of 76 messages in thread
How to write programs to internal memory with RS232???      Jacob Drake      11/11/11 00:56      
   What processor      Per Westermark      11/11/11 01:05      
      And the winner is...      Jacob Drake      11/11/11 11:13      
         fuggetabutit        Erik Malund      11/11/11 11:42      
            Hmmmmmm      Jacob Drake      11/11/11 11:53      
               sure there is      Erik Malund      11/11/11 12:15      
               Costly failure      Per Westermark      11/11/11 12:20      
                  Well, the UV thing confuses me...      Jacob Drake      11/11/11 12:26      
                     well, then ....      Erik Malund      11/11/11 12:56      
                     OTP or mask-programmed      Per Westermark      11/11/11 13:00      
                     Isn't that a clue?      Richard Erlacher      11/11/11 14:09      
                  Wrong! It won't work with THIS part!      Richard Erlacher      11/11/11 14:08      
                     8032      Per Westermark      11/11/11 14:39      
                        Actually ...      Richard Erlacher      11/11/11 23:50      
                           What situation would that be?        Per Westermark      11/12/11 05:25      
                              I think you're confusing two different parts.      Richard Erlacher      11/13/11 01:09      
                           There's the thing!      Andy Neil      11/12/11 05:37      
                              It all depends on what you mean by experienced.      Jacob Drake      11/12/11 12:54      
                                 RE: It all depends on what you mean by, "experienced"      Andy Neil      11/12/11 13:03      
                                    "Level" is such a relative term      Jacob Drake      11/12/11 17:35      
                              That's the reason      Richard Erlacher      11/13/11 01:15      
                        8032?      Jacob Drake      11/12/11 00:47      
                           Yes - but is it worth it?      Andy Neil      11/12/11 04:30      
                           It's a long, circuitous path ...      Richard Erlacher      11/13/11 01:24      
                     I have an external UV EPROM      Jacob Drake      11/12/11 00:38      
                        If you don't know that, ...      Andy Neil      11/12/11 04:35      
         Unless you have specific interest in vintage electronics ...      Andy Neil      11/11/11 13:13      
            So...I've made a decision..sorta      Jacob Drake      11/12/11 23:51      
               One step at a time      Andy Neil      11/13/11 00:21      
                  Any specific one though?      Jacob Drake      11/13/11 00:28      
                     This one?      Andy Neil      11/13/11 01:04      
         Not this one!        Richard Erlacher      11/11/11 13:55      
            Hmmm, a goal?      Jacob Drake      11/12/11 00:44      
               Wrong road!      Andy Neil      11/12/11 05:07      
                  Regarding knowledge of electronics.      Jacob Drake      11/12/11 12:57      
                     RE: "fourth grade"      Andy Neil      11/12/11 13:10      
                        RE:RE:Fourth grade      Jacob Drake      11/12/11 17:36      
               if the "self taught home-school class" refer to the archaric      Erik Malund      11/12/11 08:17      
                  When throwing out bathwater, take care to retain baby...      Andy Neil      11/12/11 10:16      
                     out of sheer nostalgia      Erik Malund      11/12/11 16:44      
                     It might, in fact, be a good thing to go back farther      Richard Erlacher      11/13/11 01:37      
                        Interesting      Jacob Drake      11/14/11 00:15      
                           You need to know much more about two things ...      Richard Erlacher      11/14/11 12:51      
                  About other processors      Jacob Drake      11/12/11 13:01      
                     The prices were on the page!      Andy Neil      11/12/11 13:07      
                     SiLabs Boards      Michael Karas      11/12/11 13:42      
                     Integrated      Per Westermark      11/12/11 14:57      
                     here      Erik Malund      11/12/11 16:34      
                        Code flash not always able to replace EEPROM      Per Westermark      11/12/11 16:56      
                           agreed      Erik Malund      11/12/11 17:24      
                              Yes      Jacob Drake      11/12/11 17:42      
                              Yes      Jacob Drake      11/12/11 17:42      
               Maybe a heretic idea...      Matthias Arndt      11/13/11 04:44      
                  There's a problem with these ...      Richard Erlacher      11/13/11 12:56      
                     Not for the intended purpose....      Matthias Arndt      11/14/11 00:49      
                        Single source        Per Westermark      11/14/11 01:02      
                           Just another chance.....      Michael Karas      11/14/11 04:46      
                        If you read his entire post ...      Richard Erlacher      11/14/11 13:00      
                           I have six other microprocessors that i found yesterday      Jacob Drake      11/14/11 13:12      
                              Well, this is not the PIC forum ...      Richard Erlacher      11/14/11 13:37      
                                 I was just saying...      Jacob Drake      11/14/11 15:02      
                              'found'stuff      Erik Malund      11/15/11 06:31      
                                 I have to agree, conditionally      Richard Erlacher      11/15/11 10:08      
                           learning about microcontrollers      Andy Neil      11/14/11 15:24      
                              Arduino is a very good choice if not going 8051 route        Per Westermark      11/14/11 16:54      
                              Careful, now ...      Richard Erlacher      11/15/11 10:38      
                     spend some $$$ on an already-assembled product      Andy Neil      11/14/11 15:14      
                        in for a penny, in for a pound ...      Richard Erlacher      11/15/11 10:57      
                           Still easier to learn to program before bringing up first hw      Per Westermark      11/15/11 11:26      
                              I'm not so sure ... maybe we'll just have to disagree      Richard Erlacher      11/16/11 11:32      
               If you're really serious about this ...      Richard Erlacher      11/13/11 14:01      
                  tools and such      Jacob Drake      11/14/11 09:58      
                     Well ... maybe your Dad can fill you in on some details      Richard Erlacher      11/14/11 13:13      
         re: 8052      Andy Peters      11/14/11 12:57      
            Switched      Per Westermark      11/14/11 13:18      
            no offense, but that's incorrect.      Richard Erlacher      11/14/11 13:24      

Back to Subject List