HP Forums

Full Version: HP-85 Error 23
You're currently viewing a stripped down version of our content. View the full version with proper formatting.

I've got another HP-85 which displays ERROR 23 : SELF TEST when turning on. I've entered ERRL {ENDLINE } and got a result of 0. I guess one or more of the RAM IC's are bad. Can anyone confirm that this diagnosis is correct?

Can I cannibalize a 16K memory module to get replacement RAM IC's, or are these IC's different?

Any help is highly appreciated!

Thanks, Juergen

Have you tried running the diagnostics from the Service ROM?

You can download the service ROM for the HP-85 from
a service training course manual from

while the service manual itself is in the MoHPC DVD collection.


PS if you do not have the ROM itself you can burn the image from the series80 site on a standard 2764 EPROM and use the 82929A Programmable ROM drawer. If you do not have an 82929A, then
check John Shadbolt's site http://www.vintagecomputers.freeserve.co.uk/hp85/prm85.htm for
the modern equivalent.


Unfortunately, I do not have access to a Service ROM. I have the Service Manual, and there I found the hint to enter ERRL {ENDLINE} after an ERROR 23 : SELF TEST. The result was 0 which indicates bad RAM IC's.

My idea was to replace all RAM IC's by the IC's of a working 16K Memory Module. But I do not know if the IC's are really the same (the labels are different). Has anybody out there ever tried that?

Regards, Juergen

Hallo Jürgen,

The problem could be with one or more RAM ICs or with a RAM controller.

You normally have 8 RAM ICs on your board. Are they socketed or soldered ?

Could you email me your faulty RAM IC references ?

I will look in my parts (basically a "parts" Hp-86B) to check if I have them.

Grüzi aus Frankreich.


Edited: 14 Nov 2005, 2:24 p.m.

I thought the RAM ICs used in the 85A were just standard 16K*1 DRAMs, as used in many other computers of the time.

I _think_ (without looking at the service manual) that they are the very common 4116 (or equivalent) chips, they certainly are in the 9915. These need 3 power supply voltages (-5V on pin 1, +12V on pin 8, +5V on pin 9), you could check these with a voltmeter (pin 16 of the RAM chips is ground) to be sure.

Hi Tony, Jürgen,

In the Hp-85 service manual, page (pdf) 210, I have only the Hp part numbers :

1MA2-002 RAM Controller

1818-1396 RAM IC

As I have sold my Hp-85, I don't know the correspondence between the Hp part nr & standard ICs.

Jürgen, can you confirm these are standard 4116 ?


Edited: 14 Nov 2005, 3:00 p.m.

Yes, the IC's are MOSTEK 4116 (MK4116J-3). The IC's in the Memory Module are labeled MB8116E. Somewhere on the Internet I found a hint that MB8116 = MK4116 so I guess it should be possible to cannibalize the Memory Module to replace the IC's in the HP-85. What do you think?

Thank you, Juergen

So this means Tony is right!

Therefore, it's not necessary to destroy you memory module! These are off the shelf components.

You can purchase 4116 Drams with standard pins readily available in an electronic store.

For example (french sorry): Mostek DRAM 4116

Mostek is just the manufacturer so I guess any standard 8 pins 4116 will do!

Price per IC is roughly 2,5 Euros


Edited: 14 Nov 2005, 5:25 p.m.

BTW, I think you should anyway follow Tony's advice to check voltages on the Chips because there is a possibility for the RAM controller to be faulty and not the Rams themselves.


Tony & Etienne, thank you much for your advice!

The service manual diagram posted a couple of messages back confirms that there are 3 supply rails to the RAMs, as I thought. There was a rarer 16K bit DRAM that used +5V only, I wanted to make sure those were not used in the 85A.

However, you should still check the power supply voltages at the RAM chips. These don't come from the RAM controller IC (which handles the logic signals), they come from the power supply board. If one is missing or incorrect you need to troubleshoot that area.

Of course the RAMs might be good, and the RAM controller IC the problem. That is an HP custom part, alas.

I hate to say this, but you really need to do a lot more tests before replacing chips. I know the 'modern way' is to replace components more or less at random until the fault goes away, but that's never worked for me. I would (at least) check all power supply outputs, the clock signals, the address and data lines to the DRAMs, etc. Only when I had a pretty good idea as to the fault would I change parts.