PDA

View Full Version : Kaypro Keyboard Konundrum



pavery
January 24th, 2010, 01:23 PM
Greetings all

Those following Tezza's Kaypro keyboard fault (http://www.vintage-computer.com/vcforum/showthread.php?18887-Kaypro-keyboard-fault/page1) may remember

On a coincidental note, a nice looking Kaypro II (http://www.trademe.co.nz/Computers/Vintage/auction-264588757.htm) has just appeared on our Trade Me site. First one Iíve seen in my three years of looking.

Well, that machine is now mine. I have only a small collection - TRS-80 Model III & a few TRS-80 Model 100's. However I want to pursue an interest with CP/M and a Kaypro should do this nicely.

Buying it was quick - getting it going however is taking time. I will do a blog soon covering the issues so far fixed - yes, I have robbed you guys of some fun. ;) However, this is the short version:

Essentially the 4164 RAM ics had very rusty pins, yet all the other ics looked great. On pulling from their sockets, many left their pins behind! These were all Mostek manufactured, around 8247 - were there quality issues for them around that time? Tezza came to the resue and loaned a full round of 4164s. Also the SIO (again Mostek) came apart - replaced that.

The thing still wouldn't boot - drives checked out, so went probing around the FDC & found the Data Separator dead. Renewed that - still no go. Then discovered the old Data Separator had corrupted tracks 0 & 1 of boot disk! Using my last boot disk with the new Data Separator - it booted! :D

Now, this next issue I am seeking advice: Keyboard Konundrum

The keyboard is in auto-repeat all the time. I've traced the fault to a port on the 8049 - pin 33 is stuck low. This port looks at a bunch of keys, but the keys are fine - there appears to be a fault within the 8049 as this pin is lo even when disconnected from the keyboard line. The port is internally pulled up to 5V - refer to 8049 datasheet. (http://home.mnet-online.de/al/mcs-48/mcs-48.pdf) I thought maybe the internal pull-up wasn't working, so I tried adding an external pull-up - thru a 61k resistor to 5V - no change - still reads low (0.1V and heads down towards zero on my DVM).

Is there any other trick I can do to keep this port Hi, when no keys are pressed? Or am destined to find a replacement 8049 - not easy and there appears to be two different types. Tezza's 8049 is different to mine, his early Kaypro II serial-numbered machine has external ROM in the keyboard.

Thanks
Philip

Chuck(G)
January 24th, 2010, 02:15 PM
I don't think there's a difference between the 8049s that you and Tezza have, other than that his pin 7 is strapped high and yours is strapped low. When high, this forces external memory to be used for program fetches. So, you can use an 8049 strapped this way to take the place of an 8039 (which has no ROM).

pavery
January 24th, 2010, 05:42 PM
I don't think there's a difference between the 8049s that you and Tezza have, other than that his pin 7 is strapped high and yours is strapped low. When high, this forces external memory to be used for program fetches. So, you can use an 8049 strapped this way to take the place of an 8039 (which has no ROM).

Hmm, I'm confused by this Chuck(G). I'm assuming the early type keyboard (with external ROM (I think it's ROM, not RAM)) has an 8049 with no code stored within - hence the external ROM. Whereas the later keyboard has code within the 8049, no external ROM. So I can't see an early version 8049 work in a later keyboard. Though I can see that a later version 8049 work in an earlier keyboard.

Am I correct that with a later keyboard, I must use a later version 8049?

Philip

Chuck(G)
January 24th, 2010, 06:00 PM
All 8049s have ROM. The ROM-less version is the 8039. I suspect that Kaypro probably used the 8049 where an 8039 was originally called for. You can jumper an 8049 to look like an 8039. Similarly, the 8749 is an 8049, but with built-in EPROM (it has a quartz window for erasing).

MikeS
January 24th, 2010, 06:07 PM
Expanding your horizons beyond Radio Shack and large Diesel engines, eh?

I'm afraid I'm confused too about what Chuck's trying to tell us...
Sounds like you might be able to replace Tez's 8049 with an 8039, or replace yours with an 8749 if you could find one and knew what was in your 8049, but that sounds kinda iffy.

Have you tried swapping with Tez's just in case?

Sure sounds like the pull-down or input buffer is shorted to ground; probably doesn't make any difference but is it being used as input or output?

Chuck(G)
January 24th, 2010, 06:23 PM
All I'm trying to say, Mike is that there is no ROM-less 8049. The ROM-less chip (otherwise identical to the 8049) is an 8039.

However...

It's possible to force the the 8049 to ignore its internal ROM and look for all respects like an 8039. Just raise the EA pin and all address references will be satisfied from external memory. My guess is that's what they're doing on Tezz's Kaypro.

Using an 8049 in this manner is also a way to re-use old, out-of-date 8049's that have already been programmed (rather than scrap them completely).

pavery
January 24th, 2010, 06:25 PM
Expanding your horizons beyond Radio Shack and large Diesel engines, eh?

I'm afraid I'm confused too about what Chuck's trying to tell us... Have you tried swapping with Tez's just in case?

Sure sounds like the pull-down or input buffer is shorted to ground; probably doesn't make any difference but is it being used as input or output?

Hi Mike, only small diesel engines :) - small bulldozer in fact. Yes, tried swapping with Tezza's. If memory serves me right (there was masses of faultfinding at Tezza's yesterday, he might be able to correct this), Tezza's 8049 in my keyboard produce several garbled chars for every keypress. It was then we noticed his board contained an external ROM, so concluded there must be different versions. Not sure if we tried my (faulty) 8049 in his good keyboard? (Tezza help me out here, please)

It's being used as an input. I thought maybe reducing the external pull-up resistor I added, to say a few k ohm to really crank some current thru there - might free it up! (Would if it was a mechanical system :))

Philip

MikeS
January 24th, 2010, 06:43 PM
Well, pretty drastic but if there are no other options I suppose it might work; I assume you mean to burn out the pull-down transistor and replace it with an external resistor? Of course it could be the FF driving it that's stuck, or the input buffer.

I assume when it's disconnected the other keys work?

pavery
January 24th, 2010, 07:15 PM
Well, pretty drastic but if there are no other options I suppose it might work; I assume you mean to burn out the pull-down transistor and replace it with an external resistor? Of course it could be the FF driving it that's stuck, or the input buffer.
I agree, more current isn't going to fix internal faults like that. I'll wait to see if anymore ideas are forthcoming before I get drastic.


I assume when it's disconnected the other keys work?
No, no. Because the fault is within the 8049 - it detects the port is active all the time - irrespective of whether it's connected to the keys. So it's full-time auto repeat. Interestingly though, pressing another key (on another port) causes the auto-repeat to pause. I assume this is n-key roll-over code taking effect in the 8049.

Philip

pavery
January 24th, 2010, 07:19 PM
It's possible to force the the 8049 to ignore its internal ROM and look for all respects like an 8039. Just raise the EA pin and all address references will be satisfied from external memory. My guess is that's what they're doing on Tezz's Kaypro.

Using an 8049 in this manner is also a way to re-use old, out-of-date 8049's that have already been programmed (rather than scrap them completely).

I getcha, Chuck(G). Arr, using up old 8049's - this makes sense. I guess it was a lot cheaper to add a ROM than buy a heap of new 8049's (which must have been quite expensive in their day).

Thanks
Philip

MikeS
January 24th, 2010, 07:35 PM
No, no. Because the fault is within the 8049 - it detects the port is active all the time - irrespective of whether it's connected to the keys.
PhilipDuh! I knew that - don't know where that question came from. That's what I get for watching TV at the same time, and it's past my bedtime (my excuse & I'm stickin' to it ;-) )

Maybe Andrew's got some ideas; IIRC he adapted a Kaypro kbd to use with his Vector Graphic box (and now that he's got the S100 PS/2 kbd adapter he won't need it any more... ;-) )

MikeS
January 24th, 2010, 07:39 PM
I getcha, Chuck(G). Arr, using up old 8049's - this makes sense. I guess it was a lot cheaper to add a ROM than buy a heap of new 8049's (which must have been quite expensive in their day).

Thanks
PhilipI got that, but how does it help? Seems to me Philip would have to add a ROM (and know the contents of the 8049 unless he could copy Tez's), not exactly trivial if, as I assume, there's no ROM socket in Philip's kbd?

tezza
January 24th, 2010, 07:52 PM
What? You mean my keyboard circuit was old and out-of-date even for it's day and hacked around to make it work with new chips?! Hurumph...:)

Hmm...I don't think we tried your 8049 IC in my keyboard Philip. From memory we tried my keyboard on your machine and my IC in your keyboard.

Tezza

Chuck(G)
January 24th, 2010, 09:13 PM
If you look at the datasheet (http://www.datasheetarchive.com/pdf-datasheets/Datasheets-14/DSA-277054.pdf), you'll see that, unlike the 8051, the 8049 has no privacy fuses, so the program can be read out. And, the good news is that 8749s (http://www.acpsurplus.com/servlet/Detail?no=17) are still available, although I don't know about your corner of the globe.

MikeS
January 24th, 2010, 10:44 PM
If you look at the datasheet (http://www.datasheetarchive.com/pdf-datasheets/Datasheets-14/DSA-277054.pdf), you'll see that, unlike the 8051, the 8049 has no privacy fuses, so the program can be read out. And, the good news is that 8749s (http://www.acpsurplus.com/servlet/Detail?no=17) are still available, although I don't know about your corner of the globe.Now that does sound like good news; if Philip has a compatible programmer he should be in business.

dave_m
January 24th, 2010, 10:48 PM
Chuck, nice find. I didn't know ACP had any.

Philip:
I live close to ACP Surplus. I'll grab an 8749 just in case. I have a Data I/O PROM burner that is supposed to handle the chip except that the Unpak2B on it needs something called a 351B-070B adapter which I have been looking for a long time. It is the same adapter needed for the 8748 chip and I have a few of those chips that I would really like to be able to program. Maybe someone else is able to program the 8749 right away??
-Dave

pavery
January 24th, 2010, 11:59 PM
Chuck, nice find. I didn't know ACP had any.

Philip:
I live close to ACP Surplus. I'll grab an 8749 just in case. I have a Data I/O PROM burner that is supposed to handle the chip except that the Unpak2B on it needs something called a 351B-070B adapter which I have been looking for a long time. It is the same adapter needed for the 8748 chip and I have a few of those chips that I would really like to be able to program. Maybe someone else is able to program the 8749 right away??
-Dave

Chuck: Thanks for sussing out the 8749 as an alternative. Wow, ACP certainly has a lot of stuff.

Dave: That's a kind offer, thanks. Especially as I don't have a EPROM programmer of any form & that I know little about programming them as well. However it sounds like the 8749 is one of the more tricky ones to program. So I presumed my first option would be to try & find someone with old Kaypro keyboards about who could pull an 8049 for me - for payment of course.

However if someone is able to burn an 8749 for me, then that would be an even better option, in terms of long term reliability. One question: how do we get the code? Does a good 8049 need to be read by the burner? I'm not certain mine is good - but we could take that chance.

Thanks
Philip

tezza
January 25th, 2010, 12:16 AM
However if someone is able to burn an 8749 for me, then that would be an even better option, in terms of long term reliability. One question: how do we get the code? Does a good 8049 need to be read by the burner? I'm not certain mine is good - but we could take that chance.


Would there be any way to get the code out of my keyboard ROM? The ROM IC itself is soldered not socketed so I imagine it wouldn't be easy to get the contents. It's not like it can be just plugged into a reader?

I wonder if anyone has dumped this keyboard ROM somewhere?

Regarding the possibility of your ROM being corrupt Philip, from our PET experience corrupt ROM code seems often be a series of repeating values. If you (or someone) did manange to read your ROM with a reader, an examination of the extracted code may reveal some clues as to whether or not it's ok?

In saying that, there must be plenty of old unused Kaypro II keyboards around. All you need is a 8049 from one that's the same as yours. Anyone reading this got a junk one they could pull an IC from?

Tez

dave_m
January 25th, 2010, 08:56 AM
The ROM IC itself is soldered not socketed
Ouch, I assumed that it was on a socket. So they have the 8049 on a socket, but not its repair ROM??

OK, then Tez's ROM can not be used. Too risky to unsolder a 30 year old 24 pin chip. Philip's 8049 ROM contents may be good if the problem is just a bad I/O input pin. The Data I/O can read it out, and hopefully other types of programmers. Perhaps a simple kludge can be built with a 40 pin socket to hold the chip and read it out and save the data as a 2K binary file.

Of course if an old keyboard can be found for a fair price, that will work too.

Chuck(G)
January 25th, 2010, 10:16 AM
Surely, there must be a Kaypro owner on this side of the Pacific who wouldn't mind lending Dave his keyboard for a short time.

Anyone?

pavery
January 25th, 2010, 11:30 AM
If you look at the datasheet (http://www.datasheetarchive.com/pdf-datasheets/Datasheets-14/DSA-277054.pdf), you'll see that, unlike the 8051, the 8049 has no privacy fuses, so the program can be read out. And, the good news is that 8749s (http://www.acpsurplus.com/servlet/Detail?no=17) are still available, although I don't know about your corner of the globe.

I note ACP Surplus also do the 8049 at $2.95, whereas the 8749 is $19.49. Would programming the 8049 be the same as for the 8749? I understand the 8749 is erasable, whereas we'd only have one shot with the 8049.

Tez (& I) are considering buying an EPROM burner anyway. However, I get the impression the 8049/8749 won't fit the average burner & would require a high-end unit. Is that correct?

Tez: Can you send me a link on the burner you were looking at, so I can investgate?

Keyboard &/or parts: I am chasing up leads on the net re obtaining another Kaypro II keyboard/8049.

I'll keep you posted
Philip

dave_m
January 25th, 2010, 12:52 PM
I note ACP Surplus also do the 8049 at $2.95, whereas the 8749 is $19.49. Would programming the 8049 be the same as for the 8749? I understand the 8749 is erasable, whereas we'd only have one shot with the 8049.

Philip, wow I had no idea that there was a PROM version of the 8049, but I checked into it and there is something called a 8049AH that looks to be field programmable. I went scrambling to check my Data I/O wall chart to see if the 8049 was listed, but alas it is not. I will have to research this further to see if it is possible to fuse this part with a Data I/O and if it uses the same algorithm as the 8749. And while I should have an 8749 shortly, I have not tracked down the adapter I need. So better keep looking for bargin keyboards.

edit: I have quickly looked at the specs for programming an 8749 and 8049AH and they seem identical, so it should be possible to program an 8049 using the pin & family code for an Intel 8749.

edit#2: The more I look into it, it looks like there is no PROM version of the 8049. What the spec sheet refers to as "programmable ROM" 8049's are really the 8749 EPROMs. I do not know why ACP shows an 8049 as that should be a mask ROM part.
-Dave

pavery
January 25th, 2010, 04:24 PM
The more I look into it, it looks like there is no PROM version of the 8049. What the spec sheet refers to as "programmable ROM" 8049's are really the 8749 EPROMs. I do not know why ACP shows an 8049 as that should be a mask ROM part.
-Dave

Thanks for that Dave. It never occurred to me that the 8049 is a "Mask" ROM. I learnt a bit about this at Tech College almost 20 years ago, but haven't used that knowledge since - and it was only the theory anyway. Sounds like the 8749 would do the job nicely & looking at the specs for Tezza's EEPROM programmer which he *may* be going to buy, it can do the 8749. (Hope you don't mind me mentioning that Tez) :blush: That's handy to know if I resort to that route.

It sure would be good be able to burn EPROMs on this side of the Pacific, particularly with the amount of vintage hardware resting up at Tezza's place!

Philip

Chuck(G)
January 25th, 2010, 05:01 PM
Just so--the 8049 was a factory mask-programmed job. You placed your order with Intel, giving them your program code and a bunch of money for setup charges and your P.O. for 1,000+ of the things and in a few weeks you got a big box full of the things--and an invoice to match.

My recollection is that you can read the program in these things with a programmer (it's been almost 30 years for me!). If not, we'll have to figure out how to punt on this one.

But I hope this explains why Tezza has a keyboard with an EPROM. Kaypro probably had a pile of the old revision of these things sitting in stock and it was far cheaper to create a new PCB with an EPROM socket than to simply put the existing stock in the landfill.

If all else fails, it might even be possible to reverse-engineer the 8049 contents using Tezza's EPROM as a guide. Keyboards, particularly those on old Z-80 boxes aren't rocket science.

MikeS
January 25th, 2010, 05:15 PM
Well, one of my programmers lists both the 8049H and the 8749, but I suppose that doesn't guarantee that it can *read* an 8049; I'll have a look around to see if anything here has an 8049 in it that I can try to read.

dave_m
January 25th, 2010, 05:15 PM
Philip,
If Tez doesn't mind, double check his selection of EPROM Programmer. If it is a new design, they don't always have the ability to program old parts because of the programming voltages required. See below:

really old parts like the 2532 need +25V
old parts like the 2764A and 8749 need +21V (and in the case of the 8749 +18V)
CMOS parts like the 27C128 need +12.5V

The vendor should be glad to supply a list of parts their machine can handle.

tezza
January 25th, 2010, 05:18 PM
Sounds like the 8749 would do the job nicely & looking at the specs for Tezza's EEPROM programmer which he *may* be going to buy, it can do the 8749. (Hope you don't mind me mentioning that Tez) :blush: That's handy to know if I resort to that route.

It sure would be good be able to burn EPROMs on this side of the Pacific, particularly with the amount of vintage hardware resting up at Tezza's place!


Hehehe...;)...pressure...pressure (*counting coins in my piggy bank*)

Yea, I probably will get one. Just emailing offline with another VCF member about this at the moment.

Tez

Chuck(G)
January 25th, 2010, 07:30 PM
Well, one of my programmers lists both the 8049H and the 8749, but I suppose that doesn't guarantee that it can *read* an 8049; I'll have a look around to see if anything here has an 8049 in it that I can try to read.

Check the innards of some old PC keyboards. Many of them used the 8048 cousin to the 8049. At least it would be good for a test.

dave_m
January 25th, 2010, 09:46 PM
Chuck,
I read where the AT keyboards used the 8049. Is this true?

tezza
January 25th, 2010, 11:31 PM
Guys,

I'm following this thread and of course I'm hopeful that Philip will soon have a 100% working Kaypro. Just letting you know that I have to bow out from the forums for a few days. I've got a business trip to Australia until Saturday (meetings and more meetings). I may not have time to visit the forums whilst over there.

However, I'm sure you'll manage without me :)

Tez

Chuck(G)
January 26th, 2010, 08:33 AM
Chuck,
I read where the AT keyboards used the 8049. Is this true?

It's been awhile, but I believe that's true, as well as for the PS/2 keyboards. I've got a couple of NCR keyboards here with 8039s and EPROMs, so it was a common setup.

dave_m
January 26th, 2010, 10:05 AM
Well, one of my programmers lists both the 8049H and the 8749, but I suppose that doesn't guarantee that it can *read* an 8049; I'll have a look around to see if anything here has an 8049 in it that I can try to read.

Mike, if the 8049H is listed by your programmer, it means it can be read (although not programmed) by your device. I'm not sure how your programmer works. Do you enter a code to identify the chip to be read or programmed? If so, is it the same code as the 8749?
-Dave

dave_m
January 26th, 2010, 06:58 PM
Reviewing the 8049 spec sheet, it is clear that the device can be read out, but with a different voltage applied to signal EA (+12V) rather than +18V for the 8749. So my programmer if given the family/pin code for an 8749 can not read the ROM data from a 8049.

Also a temporary setback: Dave Freeman of ACP sent email saying they are out of stock on the 8749.

MikeS
January 26th, 2010, 07:25 PM
Mike, if the 8049H is listed by your programmer, it means it can be read (although not programmed) by your device. I'm not sure how your programmer works. Do you enter a code to identify the chip to be read or programmed? If so, is it the same code as the 8749?
-DaveIn the immortal words of SNL's Emily Litella: "Never mind."
Don't know what I was smoking; the programmer just does the 8749 and the 8749H, no mask ROM devices.
Sorry 'bout that; gotta focus/read more closely and/or get new glasses...

Dutch
January 26th, 2010, 07:25 PM
Philip, Tezz, I've been following this thread intently. What kind of programmer are you considering purchasing anyway? I checked mine (Andromeda AR-32A) and with the correct adapter it can read the 80** family and program the equivalent 87** controller.

Even tough Philip's 8049 has a stuck input, I would think its contents would still yield a good insight to merge with Tezza's EPROM. I reconstructed a parallel card's ROM based firmware using something along these lines. It had a stuck bit as well.


Dutch

pavery
January 26th, 2010, 11:20 PM
I checked mine (Andromeda AR-32A) and with the correct adapter it can read the 80** family and program the equivalent 87** controller.

Even tough Philip's 8049 has a stuck input, I would think its contents would still yield a good insight to merge with Tezza's EPROM. I reconstructed a parallel card's ROM based firmware using something along these lines. It had a stuck bit as well.

Dutch

Thanks Dutch. We haven't finalized which programmer yet, but it's good to know the Andromeda could do it. At the moment I'm waiting to hear back re availability of a complete keyboard, sounds kinda hopeful. Otherwise, even though we're EPROM newbies, I'm sure with a bit of help from this group, we'd be able to get a good ROM build from the two keyboards.

Philip

Chuck(G)
January 27th, 2010, 08:27 AM
Alternate sources of the 8479:

http://www.hrent.com/inv.htm
Ebay item: 220534561631,180216613538

Dave, ACP is pulling your leg; see ebay 390116590300. If they don't have it, they're in violation of eBay rules.

dave_m
January 27th, 2010, 10:03 AM
Chuck,
No I don't think so, as ACP is a good outfit, I have dealt with them since the 80's. I'm guessing that in looking to fill my order, they realized they only had that ugly chip with a sticker left for the ebay offer. They have since updated their database and do not show any 8749s available on their website where I placed the order.

One important thing to note on the 8749 that I learned with all my recent research on that chip: Try not to buy the part numbers that starts with a "P", like P8749 as those units have a plastic case without a window instead of the ceramic case with a window. They are for one time programming (OTP) applications. Not good for us mistake-prone amateurs!
-Dave

Chuck(G)
January 27th, 2010, 10:26 AM
Dave, even with shipping, that seller in the Netherlands is still cheaper than ACP--and he's selling NOS.

dave_m
January 27th, 2010, 01:00 PM
Chuck,
I noticed. I wonder why so cheap when all the other vendors I've seen are between $20 and $40?

From this link below, it appears the 8749, the EPROM version of 8049 came out after the newer generation 8051/8751 family so it could not have been a huge winner and is probably rarer than the 8748. This source also states that that while the 8048 was used in the IBM PC keyboard, IBM switched to the Motorola 6805 for the AT keyboard as it was cheaper.
http://www.cpushack.net/Historyofthe8051.html

dave_m
January 27th, 2010, 02:00 PM
Dutch,
I was impressed by the specs for the Andromeda AR-32A. Although as you mentioned it would need an $80 adapter to program the 8748/8749 family. I like that it can be self booting from CD and is therefore independent of the operating system although booting from CD may not be an option with old PCs. What is the cable interface? Is it parallel port or USB? For some reason it didn't state.
http://www.arlabs.com/

-Dave

Chuck(G)
January 27th, 2010, 02:22 PM
Chuck,
I noticed. I wonder why so cheap when all the other vendors I've seen are between $20 and $40?

From this link below, it appears the 8749, the EPROM version of 8049 came out after the newer generation 8051/8751 family so it could not have been a huge winner and is probably rarer than the 8748. This source also states that that while the 8048 was used in the IBM PC keyboard, IBM switched to the Motorola 6805 for the AT keyboard as it was cheaper.
http://www.cpushack.net/Historyofthe8051.html

To the best of my recollection, the 8748/49 was never intended as a part that would be deployed out to the field on a wholesale basis. Rather, its primary use was in prototyping and initial field testing. The price differential was substantial enough between the plastic 8048 and the cerdip 8748 that you didn't want to buy the erasable version in quantity.

Given that there's no protptying being done currently for the 8048, anyone with a pile of 8748s is going to be looking at slow sales--or replacement, as in this thread's topic.

Tidbit: On impulse, I decided to open one of my circa-1987 NCR AT-style keyboards (made by Cherry in Germany) . There was an 8049 all right, and two unpopulated positions on the PCB; one for an EPROM and the other for an octal latch. There was also a jumper spot on the 8049 EA pin. So the practice of using an optional EPROM with the 8049 must have been widespread.

dave_m
January 27th, 2010, 05:03 PM
Tidbit: On impulse, I decided to open one of my circa-1987 NCR AT-style keyboards (made by Cherry in Germany) . There was an 8049 all right, and two unpopulated positions on the PCB; one for an EPROM and the other for an octal latch.
You say you saw a position for only an eight bit latch? I would think there would have to be a latch for 11 bits of address (2K). The upper 3 bits of address comes out on the lower half of Port 2 during an external instruction fetch, I wonder if those address bits are latched in the port? It's been way too long for me to remember.

Chuck(G)
January 27th, 2010, 05:18 PM
You say you saw a position for only an eight bit latch? I would think there would have to be a latch for 11 bits of address (2K). The upper 3 bits of address comes out on the lower half of Port 2 during an external instruction fetch, I wonder if those address bits are latched in the port? It's been way too long for me to remember.

Just a '373 is all that it takes. Here's a sample circuit using an 8039 (http://www.armory.com/~rstevew/Public/Micros/8048/Eds8039Board/8039.htm).

Dutch
January 27th, 2010, 07:56 PM
What is the cable interface? Is it parallel port or USB? For some reason it didn't state.
http://www.arlabs.com/

-Dave

The AR-32A is a standard parallel interface. Concerning the adapter, you don't necessarily have to buy it. The manual that they supply with the programmer contains documented schematic drawings for each one. I believe that ARLabs will sell a bare board if you didn't want to build an adapter from the ground up.

Dutch

dave_m
January 28th, 2010, 07:50 AM
Thanks for the schematic, it brought back old memories. So the upper three address lines are latched in the port saving a glue part. The MCS-48 chip family was a marvel for its time.
-Dave

Chuck(G)
January 28th, 2010, 08:24 AM
Thanks for the schematic, it brought back old memories. So the upper three address lines are latched in the port saving a glue part. The MCS-48 chip family was a marvel for its time.
-Dave

The higher-end AVRs (e.g. ATMega128, ATmega162...) do the same sort of thing with external memory. Add a '573 and a SRAM and you've got 64K of directly-addressable storage. More if you use an I/O to bank-switch. Something that PICs can't easily do.

pavery
February 18th, 2010, 01:12 PM
While I'm waiting for parts to arrive to fix my Kaypro II keyboard, I took the plunge and connected up a Radio Shack TRS-80 Model 100 to act as a stand-in keyboard. It is most successful.

You can read the full details in this blog. (http://www.classic-computers.org.nz/blog/2010-02-18-kaypro-m100-keyboard.htm)

Philip

dave_m
February 18th, 2010, 03:30 PM
I took the plunge and connected up a Radio Shack TRS-80 Model 100 to act as a stand-in keyboard.
Philip,
I think you were lucky that the Model 100 does not have true RS232 drive levels of about +/- 10V. That would not have been good for the TTL gate in the Kaypro. From what I can tell, the Model 100 has a simple driver with about +/- 3V so your Kaypro is apparently surviving OK with it although it explains why you had to invert the signal since you do not have a RS232 receiver on the kaypro keyboard interface.

What you should do is to send the signal from the M100 into the Kaypro serial port which has a proper receiver, then pick off the receiver output (U69-3) and jumper that signal to the U71-13 gate input.
-Dave

pavery
February 18th, 2010, 04:45 PM
Philip,
I think you were lucky that the Model 100 does not have true RS232 drive levels of about +/- 10V. That would not have been good for the TTL gate in the Kaypro. From what I can tell, the Model 100 has a simple driver with about +/- 3V so your Kaypro is apparently surviving OK with it although it explains why you had to invert the signal since you do not have a RS232 receiver on the kaypro keyboard interface.

What you should do is to send the signal from the M100 into the Kaypro serial port which has a proper receiver, then pick off the receiver output (U69-3) and jumper that signal to the U71-13 gate input.
-Dave

Thanks Dave, just as well guys like you are keeping us cowboys in line! I did consider incompatible Drive Levels (particularly when it didn't work), but when I bypassed the first inverter and it worked - all was forgotten (somewhat foolishly perhaps).

I shall make the change, and rewrite the blog! A bonus of your method is its mechanically elegant - I can simply plug my null modem cable into the Kaypro and just make one jumper. Very neat. New blog is here. (http://www.classic-computers.org.nz/blog/2010-02-18-kaypro-m100-keyboard.htm)

Thank you Dave

Philip

pavery
March 5th, 2010, 07:23 PM
Just to let you all know this thread has a happy ending. My quest for a replacement keyboard came up trumps with a nice guy from New York providing the goods.

My desire for a complete spare keyboard was founded as the replacement too had a few dud key switches. Using my old keyboard for spares, I now have an excellent working unit, with spares to boot.

The replacement keyboard uses an external ROM for the 8049, whereas my original uses the internal 8049 ROM. I've retained this 8049 as if the need ever arises, its ROM could be read out. (It will require the purchase of a special adapter to do this though).

Many thanks to the helpers

Philip

tezza
March 5th, 2010, 11:09 PM
Good to hear your Model III now has a desktop vintage companion Philip.

I need to continue my interrupted software collection aquisition with my own Kaypro now it too, has a working keyboard.

Tez

Al Kossow
October 29th, 2017, 10:57 AM
Reviving a long dead thread.
Has anyone ever dumped the firmware for the Keytronics Kaypro keyboards?
The keyboards I have access to are all mechanical MaxiSwitch style. Pictures and firmware
on bitsavers/kaypro
Also, it turns out the MAME folks discovered a bug in the firmware and that the kb serial port
runs a little slower than 300baud.

"Some incredibly poorly-written
code in the keyboard MCU depends on the fact that the T register is
updated during S4 then copied to A during S4 when mov a,t is executed.
If it doesn't see zero in this register when mov a,t is executed
immediately before the timer interrupt is taken, it hangs."