DLV11-J (M8043) and ECO M8043-MR002

[jmdhuse]

Hello All,

I am on a mission to track down the details of this ECO, which, as I understand it, is applied to an etch rev D to correct an IO register addressing anomaly. The information I have regarding this problem and ECO is from one of the micronote documents, which is quoted below.

I have both an etch rev D module (5013216D-P1) and etch rev E module (5013216E-P1), and I am determined to track down the details of the ECO.

Here is what I have done so far:

Using the MP-00586 schematics, I have modeled the decoding logic with a small C program and verified that the proper addresses are decoded. I have ohm'd out the decode logic of the rev D board, comparing it to the schematic, and haven't found any disagreement. This is confusing, because it suggests the schematic is the same as etch rev D, but there is not problem seen.

What I'd like help with is this: Does anyone have a rev D board that shows evidence of the ECO, which is indicated by the presence of a green ECO wire. If you have such a board, could you post a picture of the ECO wiring, so I have some idea of what the ECO is actually doing.

I know this is not a vital investigation, but what else are hobbies for? I'd like to be able to provide the information for anyone to modify their rev D board so they behave like the corrected rev E boards.

Thanks, Jon.

Here is the entry from the micronotes document regarding this problem and ECO:

DLV11-J I/O Page Address Problem Report

PROBLEM
Under certain conditions, the DLV11-J will falsely respond to bus cycles intended for other bus interface modules. When this happens, the DLV11-J address selection logic will be enabled, in addition to the address selection logic of the correct module. This results in the DLV11-J placing information onto the bus which will be OR'ed with the data from the correct module.

NOTE: DLV11-J modules at Circuit Schematic (CS) Revision E or above do not have this problem.

CONDITIONS
1) This problem will occur only when the program is performing a bus cycle that accesses the I/O page and will not occur for accesses in the 0-28Kword address space for the LSI-11.

2) This problem will occur only on DLV11-J modules that are configured to have a console port.

3) This problem is more noticeable with the 11/23 and with DMA transfers on the LSI-11 and LSI-11/2.

PROBLEMS OBSERVED
1) This problem has been noted to occur when using the bootstrap in the REV11 module to load paper tape. In this case, the REV11 ROM command "AL177560" will cause an abort to ODT. However, the command ''AL CR" will work properly.

2) Program failures and data errors may be noted when using the 28K to 30K area in the I/O page for program memory.

3) Errors may also be noted while using other I/O interfaces which have certain address bits (notably bits 5 and 8-12) that are similar to those used by the DLV11-J.

SOLUTION
Install ECO M8043-002 on the DLV11-J. This ECO brings the CS up to Revision E.

All modules processed by the Customer Repair Area (CRA) will be updated to this ECO. This update will be performed at no charge to the customer. This customer should obtain a no-charge SBA from his Sales Specialist in order to have the module updated.

QUICK CHECK
The presence of green wires or of CS Revision E or higher indicates that this ECO has been installed.

NOTE: This ECO IS REQUIRED FOR ALL DLVI1-J MODULES USED WITH THE LS1-11/23.

 

[dalmill]

I can send you cell phone camera microfiche viewer screenshots of this and subsequent M8043-S0003.

They are actually pretty readable considering.

I think the bulb may need renewing for optimum display brightness, but at least it works for now as it is probably over 35 years old!

You can DM me if interested-I am still moderated.

Dale

 

[jmdhuse]

I can send you cell phone camera microfiche viewer screenshots of this and subsequent M8043-S0003.

They are actually pretty readable considering.

I think the bulb may need renewing for optimum display brightness, but at least it works for now as it is probably over 35 years old!

You can DM me if interested-I am still moderated.

Dale

DM sent...

 

[dalmill]

Direct emailed two interested parties.

Just a quick synopsis (which I should have included previously)...

The M8043-S0002 has you drill holes in the board for pass-thru wiring to add a 1/4W 500 ohm resistor and 2N3646 transistor as well as some etch cuts.

The M8043-S0003 requires two replacement 9637 RS232 recvr chips of a minimum date code to correct the problem of "receiving data from a device with an EIA voltage swing of greater than -10 volts."

Regards,

Dale

 

[jmdhuse]

Direct emailed two interested parties.

Just a quick synopsis (which I should have included previously)...

The M8043-S0002 has you drill holes in the board for pass-thru wiring to add a 1/4W 500 ohm resistor and 2N3646 transistor as well as some etch cuts.

The M8043-S0003 requires two replacement 9637 RS232 recvr chips of a minimum date code to correct the problem of "receiving data from a device with an EIA voltage swing of greater than -10 volts."

Regards,

Dale

Ah, so here's the thing that has been confusing me... The ECO indicates that this is to be applied to CS version D to bring it up to CS version E. I have been assuming that the CS revisions correspond to the etch revisions, but now I'm not so sure. I have two etch revisions, and I was assuming that 50133216D-P1 corresponded to CS rev D and 50133216E-P1 corresponded to CS rev E, but both the D board and the E board have the resistor and transistor included in the design, not added as an ECO. As best as I can tell, the decoding logic for both of the etch versions is the same.

Can anyone explain how to correlate board etch revisions with CS revisions?

Thanks, Jon.

 

[dalmill]

The etch revision defines the design of the actual traces on the board as it came from the factory at a particular time in history.

This is typically indicated in the part number stamped (or written in solder) on the module (as you have decoded to D and E).

The etch rev will increase when corrective ECO actions result in a new module layout that eventually gets built.

CS (circuit schematic) revisions are the functionality updates, ie, adding wires and components to fix bugs.

The CS rev is usually stamped on the rear (solder side) of the module's handle(s).

A new etch rev will include previous fixes in the etch traces themselves (or sometimes not depending on complexity, cost to implement, etc).

The last page of the ECO/FCO includes a chart with both etch and CS revision history.

I don't see an etch E history in the doc I sent you, but just means I might not have or found the latest there ever was.

At least this is what my brain recalls after clearing some cobwebs.

Hope this helps.

Dale

 

[jmdhuse]

The etch revision defines the design of the actual traces on the board as it came from the factory at a particular time in history.

This is typically indicated in the part number stamped (or written in solder) on the module (as you have decoded to D and E).

The etch rev will increase when corrective ECO actions result in a new module layout that eventually gets built.

CS (circuit schematic) revisions are the functionality updates, ie, adding wires and components to fix bugs.

The CS rev is usually stamped on the rear (solder side) of the module's handle(s).

A new etch rev will include previous fixes in the etch traces themselves (or sometimes not depending on complexity, cost to implement, etc).

The last page of the ECO/FCO includes a chart with both etch and CS revision history.

I don't see an etch E history in the doc I sent you, but just means I might not have or found the latest there ever was.

At least this is what my brain recalls after clearing some cobwebs.

Hope this helps.

Dale

Yes, this helps a lot... I see a "J" mark on the etch D board handle, and an "M" on the etch E board handle. Also, on the schematics, I finally noticed a listing of the ECOs and the letters associated with each one. The MP00586 schematics list the following ECOs:

M8043-MR001 as rev D
M8043-MR002 as rev E (the one I've been interested in)
M8043-MR003 as rev F
M8043-MR004 as rev H
M8043-MR005 as rev J (reflecting the level of ECOs applied to my etch D board)
M8043-MR006 as rev K

Thanks so much for your help in understanding this.

Cheers, Jon.

 

[Radix]

Here's one for your list of DLV11-J "interesting" facts... If you refer to the "DLV11-J FILTER PANEL" schematic which I have as part of CKDLVJ1 MP00990-01 Rev B1, the pin numbering for the 40 way Berg is wrong in the diagrams shown below:

There is a note that the BERG connector is mounted on the reverse of the board - BUT the pin numbers are correct if the connector is on the same side as the DB-25 connectors...

When the connector is on the back of the board, it reverses the pin numbers 1 - 40 to run 40 - 1. It is as if they couldn't tell their PCB layout package (whatever that was in 1977!) that the connector was going to be on the other side of the board and transposed the pin numbers so it works correctly when the connector is on the back of the board.

If you use the pin numbers as shown, then pins 2 and 3 are reversed and pin 7 is disconnected at the 25-way - but turn the 40 way connector over (it is not symmetrical like the DZV/DHV connectors) and the connections are all as expected.

If you follow pin 8 (RXD) on the 10 way berg on the DLV11-J through the DEC documents, you land on pin 2 of the 25 way... but reverse the 40 way and you go to pin 3 on the opposite 25 way connector.

There's also a link missing on the drawing below (J5 27-29) and the pin numbers are in no good order either.



and on this page:

 

[Radix]

It occurs to me that this also explains why so many systems have a twist in each of the 10 way parts of the flat cable that connects to the DLV11-J ports...