NeXT
Veteran Member
I was going to go with the latches initially since I have an inventory of salvaged ones but it was taking us more than ten minutes to get the measurements and placement correct so we just made two large holes and called it good.
| VCF East | Apr 17 - 19, 2026, | InfoAge, Wall, NJ |
| VCF Latam | Apr 24 - 26, 2026, | Bahía Blanca, Argentina |
| VCF Pac. NW | May 02 - 03, 2026, | Tukwila, WA |
| VCF Southwest | May 29 - 31, 2026, | Westin Dallas Fort Worth Airport |
| VCF Southeast | Aug 01 - 02, 2026, | Atlanta, GA |
| VCF West | Aug 01 - 02, 2026, | Mountain View, CA |
| VCF Midwest | Sep 12 - 13, 2026, | Schaumburg Convention Center, IL |
| VCF Montreal V2.0 | Nov 07 - 08, 2026, | Saint-Lambert, Montreal, Canada |
| VCF SoCal | See you in 2027, | Southern CA |
-
Please review our updated Terms and Rules here
Four-Phase Systems IV/90
- Thread starter NeXT
- Start date
Chuck(G)
Gone, but never forgotten.
Oh, okay. I've also seen PCB extraction tools that hook onto the bottom of a card near the connector and are equipped with a spring-loaded handle. Hook the ends (either through holes in the PCB or card edge), squeeze the handle and pull and Bob's your uncle. Typically not cheap,, but avoids shredded fingers. Not general-purpose; custom for the application. My last exposure to one was about 40 years ago--saved my fingertips.
Al Kossow
Documentation Wizard
I would think what you would want would be an "L: that fit into the hole and a "T" on top to wrap your hand around
One for each hand, and pull..
I wonder why no one ever built any. Only useful for pulling boards.
One for each hand, and pull..
I wonder why no one ever built any. Only useful for pulling boards.
NeXT
Veteran Member
I've worked on analog switchers where it was a J-hook that was used to reach down and into holes in the board specifically for mechanically pulling but the other ends were bent into a ring you you could put a finger through and then use two of them to pull the board straight out. Central Dynamics (not General dynamics) actually had clips under the lid for storing them with the cabinet.
Al Kossow
Documentation Wizard
The absolute worst card ejectors are the cheap metal ones DG used in Novas.
With the pair of 100 pin high insertion force sockets they used, they just bent.
With the pair of 100 pin high insertion force sockets they used, they just bent.
CommodoreZ
Experienced Member
I've not had any of mine bend the lever, but rather the spring steel that they push against on the card cage frame. Or worse, the lever broke clean out of the PCB somehow... either way, it wasn't ideal.
NeXT
Veteran Member
Thank you for uploading the ForeWord docs to bitsavers Al!
This is great. ForeWord was a multi-user word processing application that ran on multiple configurations.
The docs uploaded also contain administration information for how to handle recovery of the earlier tape archive and write it back to a disk. It also contains information on fault handling, core dumping and bootstrapping which until now was missing because the reference manuals for IDOS are still missing.
This is great. ForeWord was a multi-user word processing application that ran on multiple configurations.
ForeWord word processing - Four-Phase Systems: : Free Download, Borrow, and Streaming : Internet Archive
ForeWord word processing - Four-Phase Systems
archive.org
The docs uploaded also contain administration information for how to handle recovery of the earlier tape archive and write it back to a disk. It also contains information on fault handling, core dumping and bootstrapping which until now was missing because the reference manuals for IDOS are still missing.
NeXT
Veteran Member
Two more boards came and went on ebay. I ended up wagering my paycheck and considered digging into my line of credit but ultimately it got silly by someone who from their history is sniping listings through the vintage computing section on ebay. Unlikely to be a typical gold bug. Chip collector? Either way, considering the number of people who admit they own a system you can count with one hand and none of them were bidding on this, they're gone into the black hole.
Jack, I know you are a regular here. For the love of god please reply to your emails next time. I tried all week to reach you for a private sale and got total silence. I need parts. I especially need RAM-9 chips. Uncool.
Anyways, I've scraped the listing images and sent them off to bitsavers, along with an updated chip catalog and photos of the boardset Four-Phase customized Diablo printers with to work on their console chains. I have not dumped the EPROM for the API board but I can be persuaded if someone wants to.
Jack, I know you are a regular here. For the love of god please reply to your emails next time. I tried all week to reach you for a private sale and got total silence. I need parts. I especially need RAM-9 chips. Uncool.
Anyways, I've scraped the listing images and sent them off to bitsavers, along with an updated chip catalog and photos of the boardset Four-Phase customized Diablo printers with to work on their console chains. I have not dumped the EPROM for the API board but I can be persuaded if someone wants to.
Last edited:
NeXT
Veteran Member
This is insane. I do not get it at all.....
When I started this thread in 2022, Four-Phase Systems was a complete unknown. Bitsavers didn't even have much if anything and neither did the Internet in general. I have flipped every stone imaginable to connect with whoever else is left in the world who also has any information on these machines and none of them either had systems or pockets deep enough to start considering getting back into them now so that I can tell it's still me and three other people on this planet. The last few listings it was hard to tell if I was being outsold or outbid by goldbugs but this ins just insane. We are over $3000usd ($4100cad!) and there's *someone* out there to whom the seller has alluded has bid far "closer to five figures" for this stash.
This was a seller who right now has two smaller listings for only two chips. We worked out a deal in PM's to sell it as a lot and I bid in excess of the amount it's worth for the gold to keep that crowd away. I can only speculate given how there's a crazy number of views that multiple chip collectors went in on a group buy, which means my money has no meaning here but this is very depressing because it more or less confirms that sourcing replacement chips is near-impossible on the open market.
(for those of you who want to point out that posting these prices is not helping my cause, you can shut up at any time. At least I'm keeping track of this madness.)
antiquekid3
Veteran Member
I didn't think the seller could see maximum bids from users. Has something changed?
NeXT
Veteran Member
It sounds like he was getting private confirmations of bids or something. Oh well.
Okay lets move on from that stressful tenure and go back to trying to get my machine to run using loaned parts and collected information.
MITA provided me with the I/O cables last year to duplicate and build a pinout. The blue plastic cable hoods were also a problem as they are basically NLA however another four-phase engineer offered to look into it and sent me several complete sets of 3D printed replacements which I'm VERY happy about. Meanwhile Sellam during VCF West let me look at his system to investigate a few things and get the information needed to build the missing cable that runs from TB4 on the power supply to JP3 on I/O board 2.
What was interesting with sellam's machine is the sharpie marks on all of the power supply cabling is factory. We both have it. What also I find interesting is that while MITA and Sellam have the cable up to I/O 2 mine was missing when I got it and that's why I suspected mine was missing because mine instead had three Hi/Lo switches on the back of the machine, however Sellam's did not have those switches, so now I'm wondering if this was a configuration change or option and you are not supposed to have both. There's no information on this.
(Note the wire color coding. They are all identical to Four-Phase spec but they are all flipped between terminals)
Four other things however we got at VCF West and Sellam.
-Sellam let me inspect and confirm that the Four-Phase 8503 tape drive is a Wangco 8B37 tape drive (on bitsavers the brochure calls it the Model 8)
(I am still by the way seeking a Wangco drive. It seems all of their families of drives at this time used the same 3-cable formatter interface which is not Pertec or Kennedy compatible)
-While at VCF West I played Courier for a third party. A Zero Halliburton case containing the test jig for the IV/90 processor appeared in Cupertino. I picked it up, brought it to Canada so it can be used to try and see if there is ANY life in my IV/90 after all of this insanity to make the main IV/70 cabinet is dealt with and it has been extensively photographed and uploaded to Bitsavers. Have fun!
This jig technically does not belong to me. It's on loan from its current owner who is letting me hold it until my work on this machine is done. Part of the deal was I had to take the above photo documenting of the unit.
It plugs into the memory board and the two boards that make up the CPU. It lacks a front panel to toggle anything into the processor from the jig but gives you an output display presumably in octal and three rotary selector wheels.
The controls for P and BP are total unknowns so far. There's no documentation, not surprisingly.
-I picked up one more RAM-9 chip. It's untested but like Thanos searching for the Infinity Stones, that is one last missing ram chip dropped into my otherwise complete set of ram boards. One more remains.
Here's a photo of that chip in front of Apple's DeAnza 3 office, formerly Four-Phase's second building across the street from their headquarters (now Infinite Loop)
-Someone at VCF West was selling rail kits for Model 30 Diablo drives, so now my drives can be racked! The same person was also selling a model 31 drive, albeit because of missing cover panels and a broken upper head, strictly for parts. It still has lots of good parts otherwise and a very clean HEPA cartridge.
(I still have made ZERO progress finding 8-sector packs. We may never actually end up using these in favor of the ongoing Diablo 30 emulator project.)
Lastly, on my way home from the show, RE-PC had a Documation M200 come in. I already have an M1000 but it's massive and I'm at the mercy of when the rollers fail and MITA supposedly still has decks of Four-Phase engineering cards. The M200 is a smaller machine so it might be nice to purchase a second reader so I can integrate Kyle Owen's into the machine.
However they didn't want to sell it until BossMan got the final say on price, which I wasn't going to hang around the shop all day for him to wander by, so I left my number and they said they would call when he had a number.....two weeks ago. [rolls eyes heavily]
Okay, so back to the machine.
The gerbers for the protoboard and the riser card came back so we now have the ability to prototype new boards for the standard Four-Phase backplane. The riser also has four LED's on them so you can also monitor the voltage rails.
Aaaaaaaaand we're back to the power supply again. :I
We had issues with this before and it came back on Saturday when I was smoke testing before putting the machine together and seeing if we had life which was extremely disappointing, then I got pulled away until this evening because of a work call. Sometimes the +5 comes up. More often than not you get a chirp, 0.66v, awful noises and the squawk you would get out of a crowbar circuit after the supply had been shut off and it was resetting, except this is a linear supply so that means something else is unhappy and occasionally "fixing" itself. Well the rail only has 180 ohms on it with just the empty backplane and front panel and it does not seem to be a loading issue because if we completely de-wire the system and load the +5 it still doesn't want to behave, so we will again have to remove the power supply and nail down what is going on before we can safely power the machine up.
We do not have any schematics for the power supply.
Okay lets move on from that stressful tenure and go back to trying to get my machine to run using loaned parts and collected information.
MITA provided me with the I/O cables last year to duplicate and build a pinout. The blue plastic cable hoods were also a problem as they are basically NLA however another four-phase engineer offered to look into it and sent me several complete sets of 3D printed replacements which I'm VERY happy about. Meanwhile Sellam during VCF West let me look at his system to investigate a few things and get the information needed to build the missing cable that runs from TB4 on the power supply to JP3 on I/O board 2.
What was interesting with sellam's machine is the sharpie marks on all of the power supply cabling is factory. We both have it. What also I find interesting is that while MITA and Sellam have the cable up to I/O 2 mine was missing when I got it and that's why I suspected mine was missing because mine instead had three Hi/Lo switches on the back of the machine, however Sellam's did not have those switches, so now I'm wondering if this was a configuration change or option and you are not supposed to have both. There's no information on this.
(Note the wire color coding. They are all identical to Four-Phase spec but they are all flipped between terminals)
Four other things however we got at VCF West and Sellam.
-Sellam let me inspect and confirm that the Four-Phase 8503 tape drive is a Wangco 8B37 tape drive (on bitsavers the brochure calls it the Model 8)
(I am still by the way seeking a Wangco drive. It seems all of their families of drives at this time used the same 3-cable formatter interface which is not Pertec or Kennedy compatible)
-While at VCF West I played Courier for a third party. A Zero Halliburton case containing the test jig for the IV/90 processor appeared in Cupertino. I picked it up, brought it to Canada so it can be used to try and see if there is ANY life in my IV/90 after all of this insanity to make the main IV/70 cabinet is dealt with and it has been extensively photographed and uploaded to Bitsavers. Have fun!
This jig technically does not belong to me. It's on loan from its current owner who is letting me hold it until my work on this machine is done. Part of the deal was I had to take the above photo documenting of the unit.
It plugs into the memory board and the two boards that make up the CPU. It lacks a front panel to toggle anything into the processor from the jig but gives you an output display presumably in octal and three rotary selector wheels.
The controls for P and BP are total unknowns so far. There's no documentation, not surprisingly.
-I picked up one more RAM-9 chip. It's untested but like Thanos searching for the Infinity Stones, that is one last missing ram chip dropped into my otherwise complete set of ram boards. One more remains.
Here's a photo of that chip in front of Apple's DeAnza 3 office, formerly Four-Phase's second building across the street from their headquarters (now Infinite Loop)
-Someone at VCF West was selling rail kits for Model 30 Diablo drives, so now my drives can be racked! The same person was also selling a model 31 drive, albeit because of missing cover panels and a broken upper head, strictly for parts. It still has lots of good parts otherwise and a very clean HEPA cartridge.
(I still have made ZERO progress finding 8-sector packs. We may never actually end up using these in favor of the ongoing Diablo 30 emulator project.)
Lastly, on my way home from the show, RE-PC had a Documation M200 come in. I already have an M1000 but it's massive and I'm at the mercy of when the rollers fail and MITA supposedly still has decks of Four-Phase engineering cards. The M200 is a smaller machine so it might be nice to purchase a second reader so I can integrate Kyle Owen's into the machine.
However they didn't want to sell it until BossMan got the final say on price, which I wasn't going to hang around the shop all day for him to wander by, so I left my number and they said they would call when he had a number.....two weeks ago. [rolls eyes heavily]
Okay, so back to the machine.
The gerbers for the protoboard and the riser card came back so we now have the ability to prototype new boards for the standard Four-Phase backplane. The riser also has four LED's on them so you can also monitor the voltage rails.
Aaaaaaaaand we're back to the power supply again. :I
We had issues with this before and it came back on Saturday when I was smoke testing before putting the machine together and seeing if we had life which was extremely disappointing, then I got pulled away until this evening because of a work call. Sometimes the +5 comes up. More often than not you get a chirp, 0.66v, awful noises and the squawk you would get out of a crowbar circuit after the supply had been shut off and it was resetting, except this is a linear supply so that means something else is unhappy and occasionally "fixing" itself. Well the rail only has 180 ohms on it with just the empty backplane and front panel and it does not seem to be a loading issue because if we completely de-wire the system and load the +5 it still doesn't want to behave, so we will again have to remove the power supply and nail down what is going on before we can safely power the machine up.
We do not have any schematics for the power supply.
Last edited:
pbirkel
Veteran Member
Nice connector shells! Are the design files published somewhere?
NeXT
Veteran Member
He didn't share the design files with me, however I'm sure I can ask next time I meet up with him.
NeXT
Veteran Member
This is A problem, but unfortunately not THE problem. This was on the -22v rail and on further investigation yes I forgot to trim it off but even with the PSU isolated form the rail it wasn't shorting anything.
The power supply is photographed somewhat on bitsavers already. I've been through here once already and did a quick clean and a recap of the regulator control board. I think this time around I need to spend more time determining how to power it standalone from the bench and dismantle it further to give a much deeper cleaning. All the large capacitors had already been checked for leakage but I'll probably go back through them again in case capacitance has become a problem and we are hitting a crowbar because of the inrush.
NeXT
Veteran Member
Okay yeah, the power supply needs a second pass.
Cleaning didn't reveal much beyond a bunch of stubborn dirt but the bulk filter capacitors are audibly bad. I went through these waaaay back in the thread and they passed leakage and seemed to reform and work fine but I never removed them to test and doing so now I can hear them rattle as their electrolyte has evaporated and they are completely dry, so there's at least three that need to be replaced and would absolutely explain why we were seemingly kicking the rail out. I'll take them into the lab tomorrow and better analyze them but now I get to begin the costly process to source replacements that will fit in the supply.
Cleaning didn't reveal much beyond a bunch of stubborn dirt but the bulk filter capacitors are audibly bad. I went through these waaaay back in the thread and they passed leakage and seemed to reform and work fine but I never removed them to test and doing so now I can hear them rattle as their electrolyte has evaporated and they are completely dry, so there's at least three that need to be replaced and would absolutely explain why we were seemingly kicking the rail out. I'll take them into the lab tomorrow and better analyze them but now I get to begin the costly process to source replacements that will fit in the supply.
NeXT
Veteran Member
I do believe the fault is in the fact one of the 45000mfd 15v coke cans reads completely open while two others still read good capacitance and ESR in static tests but if you shake them you can hear them rattle. So three in total need to be replaced. Digi-Key doesn't have exacts but they do have 47000mfd 25v with correctly spaced studs and can diameter for about $65 each.
The other 12000mfd cans tested fine and passed the shake test, so at least with my tools they are okay.
The other 12000mfd cans tested fine and passed the shake test, so at least with my tools they are okay.
Last edited:
NeXT
Veteran Member
Decided the hell with it. I'm replacing all six cans, plus another 25 of the little guys that are on are on all the boards.
It came to over $300 after taxes.
Minicomputers are expensive to recap.
It came to over $300 after taxes.
Minicomputers are expensive to recap.
NeXT
Veteran Member
Caps came in. Finally had a chance to take the partially disassembled power supply and give it a "better" cleaning. I probably needs better but I don't want to fully take it apart and do cosmetic things like paint the transformer because it's a lot more de-wiring. It's fine. IT cleans up with a brush, windex, a rag and compressed air.
The fun thing about speccing the new capacitors was that diameter AND height mattered. Both sides had to fit into the old clamps and both sides had bus bars. All of the new capacitors are the correct diameter but half the height.
They are also not exact replacements. 45000mfd capacitors are not surprisingly a common value, but 47000uf is. Since these are all being used as the bulk filtering of the supply we *should* be okay going up in vale on everything by one to two thousand microfarads. Same for Voltage. 15v caps were replaced with 25v spec caps because that was the closest available.
The 870uf 50v axial I cannot find any replacement for. It tests okay for ESR and capacitance out of circuit, so it will have to stay for now.
I did end up stripping out the threads on one of the new capacitor lugs while using the old screws because three wire terminals, a bus bar and assorted washers was just barely enough to get a screw to thread in. The new capacitors came with new screws that had an extra 1/8" of length so that went in and it's all good now. Next is to reinstall the regulator control board under the power supply which is mostly soldered in and a nightmare. Prior to all this as well I had no sensitivity in the +5v regulation. I want to triple-check this at the component level before putting the board back in because if it's something like a bad 1% Dale resistor, you gotta desolder and remove the board again.
The fun thing about speccing the new capacitors was that diameter AND height mattered. Both sides had to fit into the old clamps and both sides had bus bars. All of the new capacitors are the correct diameter but half the height.
They are also not exact replacements. 45000mfd capacitors are not surprisingly a common value, but 47000uf is. Since these are all being used as the bulk filtering of the supply we *should* be okay going up in vale on everything by one to two thousand microfarads. Same for Voltage. 15v caps were replaced with 25v spec caps because that was the closest available.
The 870uf 50v axial I cannot find any replacement for. It tests okay for ESR and capacitance out of circuit, so it will have to stay for now.
I did end up stripping out the threads on one of the new capacitor lugs while using the old screws because three wire terminals, a bus bar and assorted washers was just barely enough to get a screw to thread in. The new capacitors came with new screws that had an extra 1/8" of length so that went in and it's all good now. Next is to reinstall the regulator control board under the power supply which is mostly soldered in and a nightmare. Prior to all this as well I had no sensitivity in the +5v regulation. I want to triple-check this at the component level before putting the board back in because if it's something like a bad 1% Dale resistor, you gotta desolder and remove the board again.
NeXT
Veteran Member
Power cabling for the Diablo drives have been sourced.
For y'all out there being smug with your molex and D-sub connectors: Shut up. :U
Winchester connectors suck to source. It isn't so much there's a million radial pin combinations like Canon plugs which are often truly unobtanium and it's not like Winchester hardware is rare but finding that perfect in-between where you are getting the most out of what will be someone marking the hell-up out of the connector is a waiting game. These are 14 pin Female connectors, with cable hoods and threaded thumbscrews, rather than the quick release knobs. There is multiple variations on the connectors (pin diameter is in two different sizes), the pin housing (yes, they are available in male and female without pins and no, they are not interchangable with the larger or smaller pins/receivers) and the connector hood (threaded screws, quick release screws or latching side clips....with your choice of additional hood shielding for all of the above) and to get everything together into one shopping cart is hell. Case in point, the above cables with the connctors I actually need were techincally $10usd each. After combined taxes and shipping it came to $110cad...
For y'all out there being smug with your molex and D-sub connectors: Shut up. :U
Winchester connectors suck to source. It isn't so much there's a million radial pin combinations like Canon plugs which are often truly unobtanium and it's not like Winchester hardware is rare but finding that perfect in-between where you are getting the most out of what will be someone marking the hell-up out of the connector is a waiting game. These are 14 pin Female connectors, with cable hoods and threaded thumbscrews, rather than the quick release knobs. There is multiple variations on the connectors (pin diameter is in two different sizes), the pin housing (yes, they are available in male and female without pins and no, they are not interchangable with the larger or smaller pins/receivers) and the connector hood (threaded screws, quick release screws or latching side clips....with your choice of additional hood shielding for all of the above) and to get everything together into one shopping cart is hell. Case in point, the above cables with the connctors I actually need were techincally $10usd each. After combined taxes and shipping it came to $110cad...
NeXT
Veteran Member
So here's a neat little thing I spotted while browsing ebay.
These are indeed model 029 power supplies for Series 30 Diablo drives. They are for sale on ebay for $811cad and $110cad shipping (they aren't worth that much...) but you can see one of them has stickers and badging from Four-Phase. The other power supply also has a valid Four-Phase part number: 82430030
I know I mentioned before that the series 30 drives want +15v and -15v DC. Each drive peaks at 8A and settles around 4A when spun up, so really they wanted one power supply per drive. These power supplies aren't particularly hard to come by but they always come up for a few hundred dollars. They are not however all that complicated if you look at their schematic. You can find it over on Bitsavers - http://www.bitsavers.org/pdf/diablo/disk/81507-02_Diablo29_PS_Jan77.pdf
Note how +15 and -15 share a common DC return which is also tied to ground on both sides of the primary AC transformer. This makes adapting two generic power supplies as a substitute difficult because if you tried to flip the polarity of one power supply to make a negative voltage, it probably will not work or worse, you'll cause a voltage potential relative to ground upwards of 30v.
A modern solution however I think can be sourced using Meanwell supplies, as long as you pay close attention to the DC output. Correct me if wrong here.
In this case, if you buy a pair of LRS-200-15's (200w, 15Vdc, 15A) for $35 each off Digi-Key (1866-3331-ND) these supplies looking at the datasheet use isolated DC outputs. That is: the DC return is not connected to chassis ground. In theory with the positive voltage you treat + and - normally and on the negative voltage supply you can tie the + side to ground and you should be okay to use that for a negative voltage.
These are indeed model 029 power supplies for Series 30 Diablo drives. They are for sale on ebay for $811cad and $110cad shipping (they aren't worth that much...) but you can see one of them has stickers and badging from Four-Phase. The other power supply also has a valid Four-Phase part number: 82430030
I know I mentioned before that the series 30 drives want +15v and -15v DC. Each drive peaks at 8A and settles around 4A when spun up, so really they wanted one power supply per drive. These power supplies aren't particularly hard to come by but they always come up for a few hundred dollars. They are not however all that complicated if you look at their schematic. You can find it over on Bitsavers - http://www.bitsavers.org/pdf/diablo/disk/81507-02_Diablo29_PS_Jan77.pdf
Note how +15 and -15 share a common DC return which is also tied to ground on both sides of the primary AC transformer. This makes adapting two generic power supplies as a substitute difficult because if you tried to flip the polarity of one power supply to make a negative voltage, it probably will not work or worse, you'll cause a voltage potential relative to ground upwards of 30v.
A modern solution however I think can be sourced using Meanwell supplies, as long as you pay close attention to the DC output. Correct me if wrong here.
In this case, if you buy a pair of LRS-200-15's (200w, 15Vdc, 15A) for $35 each off Digi-Key (1866-3331-ND) these supplies looking at the datasheet use isolated DC outputs. That is: the DC return is not connected to chassis ground. In theory with the positive voltage you treat + and - normally and on the negative voltage supply you can tie the + side to ground and you should be okay to use that for a negative voltage.