• Please review our updated Terms and Rules here

RK05 Drive and Controller Work

glitch

glitch

Veteran Member
Joined
Feb 1, 2010
Messages
5,342
Location
Central VA
Starting a new thread, though this is related to the PDP-11/34a Cleanup and Repair thread.

I have several RK05 drives, a RK11-D backplane and board set from Conmega, and a RLV11-D that came with the MINC-23 (which is also where some of the RK05 drives are from). I'm starting into them in an effort to have a good baseline system to test with for some upcoming customer work, as well as just to get my RK05s online and usable. Ultimately, I'd like to have at least two drives working really well, and pair them with my PDP-11/10 -- one of the drives is a definite parts drive, I forget if it's had anything stolen off of it but it's in very bad shape from a mouse house.

Since I have the PDP-11/34a on the bench, and since it's got more RAM, can support a RL11 on the system backplane, etc. I will be using it to get the RK11-D up and going first, and I'll use the RK11-D to test and repair the RK05 drives. Yesterday, I got out the RK11-D backplane and cleaned it. Looks like it had a mouse mess in the A and B connectors. I picked out as much of the big chunks as I could, then washed it several times with hot soapy water, until no "particles" came out when shaking the water out. Then it got dried, and after drying the connectors were cleaned with heavy cardstock and isopropyl alcohol, followed by De-Oxit.

I mounted the RK11-D in the PDP-11/34a system chassis this morning. It is the only other backplane in the chassis, as the DD11-DK that was in there had issues back in 2017. I cleaned the four boards in the RK11-D controller set, along with a M930 terminator. Got everything installed and a Unibus jumper to the PDP-11/34a backplane, moved the 11/34a's terminator to the Unibus Out slot, and put the M930 into RK11-D slot 2 since I wouldn't be connecting drives at first (it was not 100% clear whether I needed to terminate the drive connector or not). The system came up and booted the RL02, so that's a good sign! No hung Unibus or anything.

ZRKJ?? controller diagnostics immediately blew up with wrong data errors in the RKCS, RKWC, and other registers. I started fault tracing and eventually isolated one of the bit faults to the data paths board -- bit 15 was stuck low. The other fault was harder to find: bit 6 of the RKCS, RKWC, and a few other registers was stuck low. I traced it through the printsets, replacing what I thought to be the most likely culprits first -- since it affected more than one register, I assumed it was a data paths fault. When that didn't produce results, I replaced the mux and latch on the status/control board. Still no luck! Replaced the Unibus buffer on the bus interface board...and it was still stuck low. Finally, I turned the chassis on its side and checked AF1 continuity with the Simpson 260 -- open connection! The fault ended up being a dirty Unibus Out connector pin on the PDP-11/34a system backplane. Now, that's been cleaned at least twice, so I considered it might be a break in a wire or something. Nope, just persistent crud in the connector. I used a piece of crocus cloth to polish it, and now have passed 35 rounds of ZRKJ?? without issue!

I guess the next step is to get a RK05 out and start going through it...
 
Got the Emulex UC18 Unibus SCSI controller installed, boot PROM burned, and everything set to run XXDP from SCSI (Bernoulli 90 cartridge drive), so that I don't have to move the PDP-11/34a off the bench to get to the RL02 right this minute. Currently running diagnostics to make sure the system is stable with the UC18 installed, since previous runs have been from the RL02.
 
Stable with the UC18! I swapped in the three boards from the RKV11-D and tested those with the driveless diagnostics. Passed 60+ iterations before I shut it down for lunch.
 
I went through the RK05s tonight. I have three that came with my MINC-23, and two that @Conmega traded me. Of those, three are probably repairable, and two are parts units. The two parts unit drives have very rough spindle bearings, and one of them has some pretty heavy corrosion. One of the rebuildable ones was mounted under a PC05 and now has a fair number of chads scattered throughout :p It'll need a lot of cleaning anyway, the plenum elbow has exploded into a zillion pieces.

Looks like I have two sets of rails in total, though I may have a third set packed away (I know they're Chassis Track and not for BA11 mounting boxes).

I have three scratch packs, two from @Conmega (one of which has an orange REJECT sticker on it and a note about the top cover being "bad"), and a Burroughs pack with the same hub sectoring that I think I picked up off eBay. I have a number of packs that came with the MINC-23, but they contain possibly useful data, like the routines for the MINC's I/O crate, so I don't want to run them on unknown drives, and I haven't made it to the LSSM to have Dave image them.

What's the best way to work on a RK05 -- racked, or on the bench? I probably have enough bars of lead from Fran to keep a rack from tipping! I can handle the drives on my own, and the bench they'll be on is a virtually indestructible Steelcase desk.
 
glitch said:
I went through the RK05s tonight. I have three that came with my MINC-23, and two that @Conmega traded me. Of those, three are probably repairable, and two are parts units. The two parts unit drives have very rough spindle bearings, and one of them has some pretty heavy corrosion. One of the rebuildable ones was mounted under a PC05 and now has a fair number of chads scattered throughout :p It'll need a lot of cleaning anyway, the plenum elbow has exploded into a zillion pieces.
Click to expand...
Do you think it would be useful to use an RK05 emulator configured as a tester to exercise the drives, read/write loops and seek loops, etc., or are there a sufficient number of diagnostics that run on a PDP-11 to do this?
It would be great to have fewer RK05's from becoming scrap or parts units. I'm also curious if such a tester is useful and if so, what features are needed

also... beware of the hanging chads :ROFLMAO:

glitch said:
What's the best way to work on a RK05 -- racked, or on the bench? I probably have enough bars of lead from Fran to keep a rack from tipping! I can handle the drives on my own, and the bench they'll be on is a virtually indestructible Steelcase desk.
Click to expand...
I work on my PDP-8/L on a Home Depot furniture dolly on the floor with a cushy play mat next to it to save my knees and hips. Maybe some folks prefer not to be on the floor but I find it handy to be able to flip/spin/rotate the machine easily. RK05 is considerably more heavy than the 8/L.
 
gwiley said:
Do you think it would be useful to use an RK05 emulator configured as a tester to exercise the drives, read/write loops and seek loops, etc., or are there a sufficient number of diagnostics that run on a PDP-11 to do this?
It would be great to have fewer RK05's from becoming scrap or parts units. I'm also curious if such a tester is useful and if so, what features are needed
Click to expand...

I'm not sure -- is that something that already exists? There seem to be plenty of diagnostics to run on the PDP-11, from toggle-ins, to MAINDEC tapes, to XXDP programs. I have the PDP-11/34a up and running with RL02 and Emulex UC18 SCSI for booting XXDP, and have the RK11-D controller passing driveless diags, so I think I'm in a pretty good place to start talking to drives.

gwiley said:
I work on my PDP-8/L on a Home Depot furniture dolly on the floor with a cushy play mat next to it to save my knees and hips. Maybe some folks prefer not to be on the floor but I find it handy to be able to flip/spin/rotate the machine easily. RK05 is considerably more heavy than the 8/L.
Click to expand...

The PDP-11/10 is currently on an old projector cart for this reason -- I'm done having to turn it over at the moment, and the cart makes it easy to move around, plus there's storage below it for the various bits that currently live with it, like my H765 power system test harnesses and transformer. I probably need to keep it off the floor due to our dog getting too close (she goes to work with me).

The RK05s are certainly heavy beasts, I moved four of them around the shop tonight! (the fifth is racked) I looked it up and apparently they weigh in around 110 LBS each! I will probably de-foam them away from the main work area, I assume from the looks of it that it's the crumbly kind that gets everywhere.
 
I quite literally developed a hernia from moving an RK05 up the stairs, and have been wanting to buy a rack lift to facilitate moving and installing them since then... Take good care of your back and, uh, intestines!
 
It's great you have so many drives to work with for diagnostics and testing. Be sure to lift with your legs when moving those heavy beasts around! I'd be interested to see your setup once you have a couple drives installed and working well.
 
antiquekid3 said:
I quite literally developed a hernia from moving an RK05 up the stairs, and have been wanting to buy a rack lift to facilitate moving and installing them since then... Take good care of your back and, uh, intestines!
Click to expand...

Indeed, not a tremendous weight, but definitely not goofing around weight! Fortunately, I have material handling equipment available to me, and @Conmega is going to loan me his datacenter lift when he comes through next week. That will be especially handy for *racking* the stuff, as I can manage the weight, but getting it aligned is difficult!
 
We had a lot of schematic diagrams in binders. Think DEC's field maintenance print sets. I could change out a CDC 160MB "moving head disk", about 80 pounds, by myself by just stacking them up and sitting the drive on top while I fiddled with the rails.

I doubt very many of us have that many prints lying about, but an old set of encyclopedias might do the trick.

Probably not a good idea to put drives in the upper part of the rack anyway.
 
I tend to rack them lower than the system unit for both weight distribution and ease of getting them in/out without a datacenter lift :p I see a lot of especially smaller racks where hobbyists have racked all the pack drives in the top, with the CPU chassis at the bottom, which always kind of confused me -- how are you supposed to pivot the system to get at the backplane?!

Between print sets, my grandfather's CRC handbooks, and other uncompressible tomes, I could probably manage! I shimmed the natural gas unit heater up to height using the CRC handbooks when I installed it in the shop at the house...
 
glitch said:
Between print sets, my grandfather's CRC handbooks, and other uncompressible tomes, I could probably manage! I shimmed the natural gas unit heater up to height using the CRC handbooks when I installed it in the shop at the house...
Click to expand...
I have a variety of scrap 2-by lumber that I use to crib my way up, It's slow going at ~1.5" per layer but starting with a layer of 1 gallon paint cans fairly easily gets me to a second 6U position from the bottom. Which is as far as I go using cribbing. And the 1.5" lift isn't burdensome. Same solution for dealing with replacing our over-range microwave. Slow and steady winds the race ... and saves the back. When you don't have a helper set of extra hands ...
 
I'd originally intended to use our high-bay shop for jobs like this, since it's big and open and I can drive the forklift in, which obviously makes tasks like this trivial :p It's currently full of pick-and-place machines though...
 
Moved the DEC half-rack with the RK05 @Conmega racked into the front repair area, next to the H960 with the MicroPDP-11/83:



The drive itself:



There's a sprinkling of chads through it from the PC05 mounted above it! First steps are obviously foam, head retract batteries, and reforming the power supply caps.
 
glitch said:
There's a sprinkling of chads through it from the PC05 mounted above it! First steps are obviously foam, head retract batteries, and reforming the power supply caps.
Click to expand...

The drive works fine without the batteries. I found foam weather strip at Home Depot that was a good replacement for the blower and the drive-to-pack.
 
What are folks doing for filters these days? I have three or four RK05s that I'd love to get back online as well. I hope my RK8E is still passing diskless diagnostics...
 
m_thompson said:
The drive works fine without the batteries. I found foam weather strip at Home Depot that was a good replacement for the blower and the drive-to-pack.
Click to expand...

Yeah, I'll probably run it with the head load disabled to purge for a while, with the batteries out. I won't be running without them and a pack mounted due to the higher risk of head crash if the power fails.

Weather strip was my plan as well, I've heard from several folks it's fine.

antiquekid3 said:
What are folks doing for filters these days? I have three or four RK05s that I'd love to get back online as well. I hope my RK8E is still passing diskless diagnostics...
Click to expand...

I'm using washable filter foam intended for windows air conditioner units from the hardware store for the intake filter. I actually have some spare HEPA filters, but will see if I can clean the ones in the drives by reversing airflow with the shop vac, as recommended on the Internet.

Roland Huisman said:
I've had some adventures with 5 RK05 drives over here. Maybe you find my postings useful...

Regards, Roland
Click to expand...

Thanks, I'll read through it!
 
antiquekid3 said:
What are folks doing for filters these days? I have three or four RK05s that I'd love to get back online as well. I hope my RK8E is still passing diskless diagnostics...
Click to expand...
There are two different HEPA filters for RK05 drives depending on their age. I bought a few NOS filters about 10 years ago.
 
re: comments regarding the need to replace the RK05 head-retract battery pack. For this type of application does it need to be a battery pack? Could a 'supercap' be just as effective? Would the RK05 battery charging circuit need extensive mods to work with an appropriately sized supercapacitor?
 
You must log in or register to reply here.
Back
Top