HP 71B Diagnostic ROM

[Dave Frederickson]

I recently bought a 71B and memory module off EBay. Does anyone have a diagnostic ROM I could borrow (or buy) to thoroughly check it out?

Dave

[Reinhard Hawel]

Hi Dave,

you really don't need the Diagnostic ROM except you want to repair a 71B. Usually you'll only need to Press

[ATTN] (=ON) and [ / ] simultaneously.

When you see the String "Init: 1" push the "2" key and then press [End Line]

The Rom Test should Show a "G" (for Good) after the Digits 1 2 3 4 (for the four ROMs) if you see a "B" you have a defunct ROM (which is really bad, I think).

This should be sufficient, because I think, when your 71B works satisfactorily, you won't need any more test functions.

BTW: Here's a Link to the Selftest Functions of some (all?) HP Calculators:

http://www.rskey.org/~mwsebastian/selfte....htm#hp71b

Postscript: If anyone has a Diagnostic ROM left, I'd also be interested :-)

[Dave Frederickson]

Hi Reinhard,

Thanks for the suggestion, but I don't think any of the 71's self tests check RAM.

Dave

[Raymond Del Tondo]

Hi Dave,

you could write a small BASIC or Saturn assembly program to check the RAM cells.
Given the relatively limited amount of maximum RAM such a program should get through within a few minutes.
While we're at it: Did you check if such a program already exists, maybe on the swap disks?

[Reinhard Hawel]

Hi Guys,

You could separate the memory module(s) into Ports and check them module by module. The 71B contains four memory "modules" (in fact the internal memory consists of some small RAM zone plus four chips identical to a memory module.
Separating these modules into ports will keep your machine alive without lock up when you run programs. The small 1.5k portion has to be tested separately (and very careful while a program is running). I'm sure, when you can execute even small BASIC programs this is a sufficient test for the 1.5k. You could DIM an ARRAY and fill it with numbers or strings and read that back. Unfortunately this is far from exact for a memory test.

How to test a port:


100 COPY FILE1 TO PORT(0.01) ! 0.01 is only an example. Execute SHOW PORT for more locations
110 A = IP(ADDR$("FILE1")/256)*256 ! Beginning of the Module
120 P1$ = "AAAAAAAAAAAAAAAA" ! Test Pattern 1 16 Times 1010b
130 P2$ = "5555555555555555" ! Test Pattern 2 16 Times 0101b
140 FOR I = 1 to 256 * 2 ! 511 times 16 nibbles, first 16 nibbles not tested
150 A$ = DTH$ ( A + I * 16 ) ! Start Address as string
160 POKE (A$, P1$) ! Poke Test Pattern 1
170 T1$ = PEEK$ ( A$, 16) ! Read memory
180 POKE (A$, P1$) ! Poke Test Pattern 2
190 T2$ = PEEK$ ( A$, 16) ! Read memory
200 IF T1$ # P1S OR T2$ # P2$ THEN DISP "Mem fail at"; A$ @ I = 256 * 2
210 NEXT I

FILE1 stands for an arbitrary file name. Take any file you like.
I do not check the first 16 nibbles in line 140 because I believe, that there is some sensitive data for the file chain in the port. I think POKEing there is forbidden, but I'm not sure about that, because I write these lines without an 71B around me and no way to test it now. So please apologize my bad 71B BASIC. After all, I worked with this machine 30 - 20 years ago and this might be my first HP71B BASIC program after 10 years.
Besides that, I'm aware of the fact, that this is a really poor way of memory testing, but it's 99% sufficient and easy to write.

I'd be glad about comments

Reinhard

[Reinhard Hawel]

Hi everybody,

(12-30-2013 07:03 AM)Raymond Del Tondo Wrote:  Hi Dave,

you could write a small BASIC or Saturn assembly program to check the RAM cells.
Given the relatively limited amount of maximum RAM such a program should get through within a few minutes.
While we're at it: Did you check if such a program already exists, maybe on the swap disks?

I don't remember such stuff on the swap disks (without looking there now). HP 71s failed rarely in the old times and in fact I've never seen a CPU, RAM or ROM failure

[Dave Frederickson]

Thanks everyone for your suggestions. This isn't the first time the topic of memory testing has come up and I've read many of the old posts. I've been intrigued by the Diagnostic ROM since I first learned of it decades ago. The acquisition of a new 71 and memory module seemed like a good opportunity to see if one was available for loan and I am interested in performing a thorough test.
In addition to the Diag ROM, HP also had a "71 Service ROM Hybrid 4-Int", P/N 00071-60908 ($32). Anybody know what this was for?
Speaking of this newly acquired 71, it's the newer version without the flex connecting the PCBs, with date code 2707, yet the ROM version is 1BBBB. There's a 2CDCC on EBay with a date code of 2515. Why does this newer 71 have older firmware?

Dave

[Dave Frederickson]

Yes, I see the Diag ROM on TAS, Item # 191133764900. Only $1450 USD!

[Raymond Del Tondo]

(12-30-2013 06:53 PM)Dave Frederickson Wrote:  Speaking of this newly acquired 71, it's the newer version without the flex connecting the PCBs, with date code 2707, yet the ROM version is 1BBBB. There's a 2CDCC on EBay with a date code of 2515. Why does this newer 71 have older firmware?

I'm sure the HP-71B with datecode 2515 either has transplanted innards from a later unit, or at least the back side has been replaced.

I have seen HP-71B units from 1987 with ROM 1BBBB, and the update to ROM 2CCCC happened in 1987, and shortly after the first upgrade the ROM 2CDCC was released.

[Dave Frederickson]

Is that likely? Most 1BBBB 71's have a flex cable connecting the two PCB while the 2CDCC's have a plastic back and are connected by a "golden zebra", like the one pictured in the FRAM71 documentation.
When the time comes to sell one of my 71B's I'll just call it "rare".

[Joe Horn]

(04-13-2014 01:24 AM)Raymond Del Tondo Wrote:  I'm sure the HP-71B with datecode 2515 either has transplanted innards from a later unit, or at least the back side has been replaced.
I have seen HP-71B units from 1987 with ROM 1BBBB, and the update to ROM 2CCCC happened in 1987, and shortly after the first upgrade the ROM 2CDCC was released.

Cool, I've got a '71B Frankencalc: Date code 2412, VER$ 2CDCC, metal back.

[Dave Frederickson]

I have a theory. First, here's my 1BBBB from '84. Notice the brightly colored key legends and the three screws visible on the back side. This is typical of the early models with 1BBBB ROM versions and the metal back. Date code 2422A.



Next is a 2CDCC with a plastic back. Notice the lighter colored key legends and the absence of the screws in the back. Also, there are two new screws in the battery compartment, located near the "golden zebra" connector. See the FRAM71 doc for a picture of the board-to-board connector. Date Code 2902A





Finally, here is the 71B in question. Date code 2707A.


I don't see how an old front and a new back, or vise versa, could be possible since the connector is physically different. The old version has a fragile flex cable and the newer version a zebra connector. The interior shows no signs of refurbishment, unlike my 2422A which I had to send in for repair. HP left a little note stuck to the top of the HP-IL compartment.
I suggest that when the CPU/Display board was being assembled that 1BBBB ROMs were used, perhaps because HP's disposition of old version ROMs was to use until depleted.

My explanation for Joe's Frankencalc is that at one time it had a lobotomy. Someone repaired the calculator by replacing all the ROMs. Joe, does your 71 have bright or light colored key legends?

Dave

[Joe Horn]

(04-13-2014 07:35 PM)Dave Frederickson Wrote:  My explanation for Joe's Frankencalc is that at one time it had a lobotomy. Someone repaired the calculator by replacing all the ROMs. Joe, does your 71 have bright or light colored key legends?

It has the same brightly colored key legends as my 1BBBB machines and your top photo above. It also has only one screw in the battery compartment, not three, same as the 1BBBB machines. However, its battery terminals have foam pads (like in your bottom photo), not springs (like my 1BBBB machines have):



So it looks like the change from battery compartment springs to foam pads was not simultaneous with the change to plastic backs. IMHO it was a stupid change; springs last forever, but foam deteriorates and so does the glue that holds the little metal contact plates onto the foam. But I digress.

[Dave Frederickson]

That's makes sense. My 1BBBB went back to HP for repair. They found corroded battery terminals and a cracked flex cable. The spring terminals were replaced with the foam terminals like in the pic, above. Your 71 received new terminals and new ROMS. I'll bet that if your 71 was opened we'd find a little note from HP stuck to the HP-IL compartment, just like mine.

The cool thing about the repair of my 71 is that Duracell paid the $150 to get it fixed because it was their batteries that leaked!

Dave