I forgot to mention that I did find that my RD53 was configured as DS3 and changed it to DS1 (numbering from 1-4, not 0-3), but that made no difference. I tried swapping the jumper on the M9058 from the factory setting of TP1-TP2 to TP2-TP3, but it made no difference. Is that the jumper you mean? Regards Rob pin? for won't DS2 select so select and J4 and to pin and go

of links. cut/join for tester I used my digital multimeter to check the pins you mention, there is continuity between pins 26, 28, 30 and 32 on each of J1 to J4. I also checked that pin 4 on J9 is connected to pin 12 on J12, this is the RD 0 RDY H signal, so the drive should be seen as ready by the RQDX3. I have also checked that J5 pin 1 is connected to J12 pin 14 which is the select ack (DRV SEL 0 ACK (L) ), so any selection confirmation should be coming back to the RQDX3. I suspect I need a way to monitor the signals actually passing across the wires, I have a multimeter and an oscilloscope, but I am not sure of the practicalities of how to get access to the signals, I suppose I need some kind of "breakout" ribbon cable, but I don't want to ruin the existing cables. Is there some standard technique for doing this kind of work? For now I think I will just solder some diagnostic wires directly to the M9058. I have put it back to DS3, although it is my understanding that with an M9058 it does not matter what setting you use. I just checked, and yes it does.

On 01/09/2011 12:28, Rob Jarratt wrote: So that's OK. It used to be possible to get a little adapter that goes in-line with a 34-way (other sizes also available); it consisted of a rectangular block with one female contacts on one side, male pins on the other, and a second set of male pins on top, to which one would attach probes. They were stupidly expensive, and I'm not parting with mine :-) However, you can achieve exactly the same effect by mounting two male IDC connectors and one female on a short piece of ribbon cable. Or add a male IDC to a short female-to-female so you can use it as an extension and make two or three probe cables with square wire-wrap pins -- those are an exact fit in a female IDC connector. Or even just add one more female IDC to the existing cable and probe that with wire-wrap pins. The pins I used for mine are gold-plated single wire-wrap IC socket pins, but anything similar will do. Sounds like the software has a problem rather than the hardware. I looked back at earlier posts; you said you tested the RD53 in a MicroVAX 2000, and the RQDX3 in a BA23 uVAX II. But did that test use the same RQDX3 with the RD53 in the uVAX II? I /think/ the format put on an RDxx by any revision of RQDX3 is the same, though new firmware might change things like RCTs in ways not transparent to old firmware. But the format used by RQDX1 and RQDX2 is quite different to RQDX3 and I wonder if the drive is the "wrong" format? Isn't a uVAX 2000 the same format as RQDX1/2? The last possibility I can think of is that there are a few active devices on an M9058. They're mostly Schmitt triggers and open-collector inverters used to buffer signals from RQDX to drive (and v.v.), but some are differential buffers for the MFM data. Maybe one of those has gone. You could of course test the M9058 in the BA23. Just ignore the BA23's normal drive cabling and connect the RQDX in the BA23 via the 50-pin cable to the M9058, and cable the drive to that as if it were in a BA123. The M9058 has no connections to the backplane except for power and won't interfere with anything else. -- Pete Peter Turnbull Network Manager University of York

low-level unstuck readable once Just tried this, I put the RD53 that won't work into the BA23 and it worked fine. So I think the M9058 must be faulty. I will see if I can identify a faulty component on the board. one Further thought on this suggestion is that it would be physically difficult to achieve this because the M9058 has to be in a Qbus slot and it would be impossible to connect the RD53 to the connectors as the cables I have are way too short. I am not convinced that this test is worth the effort of making some cables to fit because I have already eliminated everything except the M9058, I think my best bet now is to check for faulty components and see if I can buy or borrow a spare M9058. hoping to one. Regards Rob

not so OK, but given all the other combinations I have tried I didn't think that trying the RQDX3 in the BA23 would make any difference, anyway I tried it all the same. To confirm, I put the RQDX3 and the RD53 from the BA123 where the disk is seen as offline, into the BA23, essentially testing with a different distribution panel and a different cable from the RQDX3 to the panel. In this configuration the disk was seen and the MicroVAX II booted off it just fine. Just for good measure I also took the cable from the RQDX3 to the distribution board from the BA123 and put that in the BA23. So now I have the RQDX3, the RD53, and the cable from RQDX3 to the distribution board, all taken from the BA123 and put into the BA23, so now testing just with a different distribution panel (NB I have already tested the cables from the distribution panel to the RD53). It still sees the disk and boots off it. This means it just has to be the distribution panel or the front panel switches. But I checked the Ready switch and observed at the output of the distribution board and at the input of the RQDX3 that the switch open and closed the contact. I think all that is left is the distribution panel itself. What I need now is to borrow or buy an M9058 to test with, and ideally a qbus extender so I can use an oscilloscope to work out which component is failing. Regards Rob

It may be the chip, but what are the control inputs doing? Pin 15 is a global enable to that chip, if it's high, the chip is disabled, and all ouptus are forced high (IIRC). Often this is not used, and pin 15 is tied to ground, but it's best to check. Pin 1 selects between the 1 possible inputs for each multiplexer section. Is this in the right state to pass the data through? -tony

What about pin 1? I've not looked at the printset, so I don't know how the '158 is used here, but pin 1 is the select input. If it's low, pin 4 is the inverse of pin 2 (and pin 3 is ignored), if it's high, pin 4 is th einverse of pin 3 (and pin 2 is ignored). The expression 'GIGO' applies to logic gates (and other devices) as much as to complete computers :-). If the input conditons aren't right, the output won't be right either, even if the chip is working perfectly. -tony

Often DEC-numbered parts are selected versions of standard ICs with the same number, bnd the parameter they're sleected for may not really matter (!). Have you tried puting '8837' in one of the data sheet search engines to see what comes up. If the function and pinout looks sane, it may be a suitable device. Hw many pins does htis device have? What is the full functionality (or can you point me to an on-line copy of the printset so I can look over the schematic)? In particualr, does this cjip have enable inputs, and are they used (that is, not tied to +5V/gnd as appropriate)? Yes, you really do have to check everything. -tony

allison wrote about the DEC8837: junk interfacing DEC didn't develop it, National Semiconductor developed it, and sold it as the DS8837. DEC had them binned for slightly tighter receiver threshold and leakage current specifications, so in principle a DS8837 might not meet the DEC specifications for a DEC8837. In practice they almost always did, as National Semiconductor was fairly conservative about the specifications for the standard part.

allison wrote: Given that National didn't EOL them until 1998, and the DS8641 and DS3662 not until 1999, I don't think that's true. There must have been other users. National had app notes for them pushing general-purpose use, with nary a word about DEC.. If DEC hadn't been consuming them, they might have gone away sooner. The non-DEC uses lasted longer, but all the things that those were used for eventually got replaced by more modern signalling standards.

"RECEIVER, BUS, to suitable you schematic)? is, not everything. The PDF of the printset came from here: http://bitsavers.org/pdf/dec/qbus/MP02259_RQDX3_schem_Oct85.pdf Look at p12 for the M5098, the component that is not working is E1, specifically pin 3 is high and pin 4 is high, but pin 1 is high and pin 2 is low. I checked the enable inputs too, pins 7 and 9, and they are both low, which looks correct to me. There seems to be a National Semiconductor chip with a similar designation and description (http://www.datasheetcatalog.org/datasheets/120/216865_DS.pdf), but that does not seem to be made any more and I don't know what the equivalent is (if any). Someone has offered me a couple of 8837s, but if I can find a source of equivalents I would prefer to buy those rather than use someone's spares. Regards Rob

thing to Eric, Your other reply suggests that the DEC 8837 does have an almost equivalent in the NS DS8837 which I have found a source for. I am unclear if you think NS devices would do or not? What is the worst that could happen if the NS part was not up to DEC specifications? Thanks Rob

equivalent think it should work. specifications? A system near the maximum configuration limits might be less reliable. For a more typical system there would be no issue.

(but you'd disk I am going to go with the "standard" part. If the disk is corrupted I will be able to rebuild it. I am more interested in the likelihood of permanent physical damage to other parts, but that seems unlikely. Regards Rob

I would too. I can't believe you'll ahve any problems/ I really can't see how it could damage anything else. It's not going to draw excessive current (either from the inputs or the power rail). -tony

the to is low. which Just after sending the above I thought I would check at your favourite (Cricklewood Electronics) and also eBay, I found some 8837s there, so I think I will order these. Regards Rob

will the I will put it back to DS3, but it didn't work when set that way. My understanding of the docs is that the DS1-4 jumper is irrelevant when using an M9058. problem sure. The M9058 has to be the focus for the moment as it is the only bit I have not been able to test separately. RQDX2 or which hard the At the moment I only have one drive in the machine, although I will be adding a second one once I have cleared up this problem. That said, this is not a BA23, but a BA123 and as I said above, I believe the M9058 handles drive selection so that the DS jumper settings are irrelevant. pins due Yes, I am aware of the problem with the heads getting stuck. I have cleaned out the sticky bumper and verified that the disk works in a MicroVAX 2000. I am about to repeat the test in the 2000 just to be sure. However I am also pretty sure the heads are not sticking because I don't hear the drive spinning down until I actually switch the machine off. Thanks Rob