Re: IBM transistor replacements
On 11/27/21 2:34 PM, Al via cctalk wrote: Thank you all for your responses and the leads you have given, Responses to some of your questions and some more questions: "Have you checked/measured whether they're actually faulty?" Yes I have, in the picture I have provided on the bottom there are the ends of 3 transistor leads. This was one that fell off as I removed the card and by far the worst corroded, so despite the fact that the rest still work. I need to replace this one. Yes, saw that in the pic. Maybe you can scrape the leads and solder a blob to that one and take some measurements, if there is enough lead length left. First thing is to determine if it is Germanium or Silicon. The similar leads rusted through issue sounds like these may be from a transitional period between SMS-Germanium and SLT-Silicon, so they might be Ge. If so, the forward drop of the BE junction will be about 0.3 V, instead of the Silicon's 0.7 V. Finding Ge transistors today will not be easy. Jon
Re: IBM transistor replacements
139 is a "nickname" number. IBM used these nicknames for some reason, instead of the full IBM part number. A 139 is really an IBM part number 2414938, a silicon device in a TO-18. Gee, it would be great if this list could take pictures so I could just post one with all the specs... -- Will On Sat, Nov 27, 2021 at 5:54 PM Brent Hilpert via cctalk wrote: > > On 2021-Nov-27, at 11:22 AM, Gregory Beat via cctalk wrote: > > > > The Texas Instruments (TI) 139 is likely a 2N139 PNP (BJT) transistor, > > capable of high speed switching (in that era). > > > > The 2N139 was originally an RCA transistor (tall cylinder) found in RCA and > > GE transistor radios (455 kHz IF section). > > The TI versions were low profile metal case, TO-33 case (8.5 to 9.5mm > > diameter) > > > From the JEDEC number, 2N139 would be a 50s-era transistor, way earlier than > the ones at issue. > 2N139 specs are Ge-PNP, 16V/12V CE max, 15mA IC max, 35mW max, which can find > a use in transistor radios, but highly unlikely to be what the OP is looking > for.
Re: IBM transistor replacements
On 2021-Nov-27, at 11:22 AM, Gregory Beat via cctalk wrote: > > The Texas Instruments (TI) 139 is likely a 2N139 PNP (BJT) transistor, > capable of high speed switching (in that era). > > The 2N139 was originally an RCA transistor (tall cylinder) found in RCA and > GE transistor radios (455 kHz IF section). > The TI versions were low profile metal case, TO-33 case (8.5 to 9.5mm > diameter) From the JEDEC number, 2N139 would be a 50s-era transistor, way earlier than the ones at issue. 2N139 specs are Ge-PNP, 16V/12V CE max, 15mA IC max, 35mW max, which can find a use in transistor radios, but highly unlikely to be what the OP is looking for.
Re: IBM transistor replacements
Greg, Thank you, I will see if I can find a modern day equivalent of that and compare them, always useful to have spares on hand. Thanks, Al (Also, apologies for the hideous formatting on my last email, something got very broken somewhere) > On 27 Nov 2021, at 21:00, Gregory Beat via cctalk > wrote: > > Al - > > The Texas Instruments (TI) 139 is likely a 2N139 PNP (BJT) transistor, > capable of high speed switching (in that era). > > The 2N139 was originally an RCA transistor (tall cylinder) found in RCA and > GE transistor radios (455 kHz IF section). > The TI versions were low profile metal case, TO-33 case (8.5 to 9.5mm > diameter) > > greg > == > From: Al > To: cctalk@classiccmp.org > Subject: IBM transistor replacements > > Dear all, > > A while ago I received an IBM 3286 printer, annoyingly some of the > transistors in the printer section have been corroded. > What I am having trouble with is reading the part codes and finding a modern > equivalent of them. > > One has a Ti logo and two sets of numbers (attached). Does anybody know which > numbers are the part numbers and if they are IBM house numbered? > > Photos of the Ti transistors and card assembly: > https://drive.google.com/drive/folders/1JTnsWS4A8NrYJNFsqK2ejOeqqJNjoF1w?usp=sharing > > Thanks in advance, > Al
Re: IBM transistor replacements
Apologies for the previous email and it’s disgusting formatting, here is a second attempt, hopefully it’s readable. Thank you all for your responses and leads, Responses to some of your questions and some more questions: "Have you checked/measured whether they're actually faulty?" Yes I have, in the picture I have provided on the bottom there are the ends of 3 transistor leads. This was one that fell off as I removed the card and by far the worst corroded, so despite the fact that the rest still work. I need to replace this one. “I would suspect that these transistors are pre-drivers for the power transistors and may not be super critical to get an exact replacement." Would an NPN or PNP with the same HFE suffice, or do I need to know more? Forgive me but I am a complete idiot when it comes to this kind of circuit design and their tolerances. "Did you get manuals with the printer? If so it should have ALDs (schematics)." Unfortunately not, and nor can I find any online. "The 3286 printers and 3277 terminals where infamous for intermittent problems" Excellent, glad I bought something that will provide the true IBM experience. Thank you, Al > On 27 Nov 2021, at 20:35, Bill Gunshannon via cctalk > wrote: > > On 11/27/21 3:17 PM, Van Snyder via cctalk wrote: >>> On Fri, 2021-11-26 at 22:33 -0800, Brent Hilpert via cctalk wrote: On 2021-Nov-26, at 1:52 PM, Al via cctalk wrote: >>> A while ago I received an IBM 3286 printer, annoyingly some of the transistors in the printer section have been corroded. What I am having trouble with is reading the part codes and finding a modern equivalent of them. There are two types. One has a Ti logo and two sets of numbers (attached). Does anybody know which numbers are the part numbers and if they are IBM house numbered? Photos of the Ti transistors and card assembly: https://drive.google.com/drive/folders/1JTnsWS4A8NrYJNFsqK2ejOeqqJNjoF1w?usp=sharing >>> >>> >>> >>> Have you checked/measured whether they're actually faulty? Yes, they >>> look terrible, but that doesn't mean they're faulty. It looks like >>> minor corrosion of the steel case under its plating; unless it's all >>> the way through the die is probably fine and cozy inside. The >>> electrically active parts of the transistor typically aren't as >>> susceptible to corrosion as the steel case is. >>> >>> (Notable exception to this being 70's-era TI ICs with plated steel >>> pins. I've also seen some Motorola ICs with corroded pins). >> The Computer History Museum in Sunnyvale, CA has two IBM 1401's that >> are in operating order. One of the problems they had in getting one of >> them to work was that some of the IBM transistors also have plated >> steel leads that had rusted through. Ask Rob Garner , >> the leader of the 1401 restoration project, if he recognizes the >> numbers. > > Make sure you keep the room air conditioning in working order if > you want an operating 1401. :-) > > bill >
Re: IBM transistor replacements
Al - The Texas Instruments (TI) 139 is likely a 2N139 PNP (BJT) transistor, capable of high speed switching (in that era). The 2N139 was originally an RCA transistor (tall cylinder) found in RCA and GE transistor radios (455 kHz IF section). The TI versions were low profile metal case, TO-33 case (8.5 to 9.5mm diameter) greg == From: Al To: cctalk@classiccmp.org Subject: IBM transistor replacements Dear all, A while ago I received an IBM 3286 printer, annoyingly some of the transistors in the printer section have been corroded. What I am having trouble with is reading the part codes and finding a modern equivalent of them. One has a Ti logo and two sets of numbers (attached). Does anybody know which numbers are the part numbers and if they are IBM house numbered? Photos of the Ti transistors and card assembly: https://drive.google.com/drive/folders/1JTnsWS4A8NrYJNFsqK2ejOeqqJNjoF1w?usp=sharing Thanks in advance, Al
Re: IBM transistor replacements
On 11/27/21 3:17 PM, Van Snyder via cctalk wrote: On Fri, 2021-11-26 at 22:33 -0800, Brent Hilpert via cctalk wrote: On 2021-Nov-26, at 1:52 PM, Al via cctalk wrote: A while ago I received an IBM 3286 printer, annoyingly some of the transistors in the printer section have been corroded. What I am having trouble with is reading the part codes and finding a modern equivalent of them. There are two types. One has a Ti logo and two sets of numbers (attached). Does anybody know which numbers are the part numbers and if they are IBM house numbered? Photos of the Ti transistors and card assembly: https://drive.google.com/drive/folders/1JTnsWS4A8NrYJNFsqK2ejOeqqJNjoF1w?usp=sharing Have you checked/measured whether they're actually faulty? Yes, they look terrible, but that doesn't mean they're faulty. It looks like minor corrosion of the steel case under its plating; unless it's all the way through the die is probably fine and cozy inside. The electrically active parts of the transistor typically aren't as susceptible to corrosion as the steel case is. (Notable exception to this being 70's-era TI ICs with plated steel pins. I've also seen some Motorola ICs with corroded pins). The Computer History Museum in Sunnyvale, CA has two IBM 1401's that are in operating order. One of the problems they had in getting one of them to work was that some of the IBM transistors also have plated steel leads that had rusted through. Ask Rob Garner , the leader of the 1401 restoration project, if he recognizes the numbers. Make sure you keep the room air conditioning in working order if you want an operating 1401. :-) bill
Re: IBM transistor replacements
Thank you all for your responses and the leads you have given, Responses to some of your questions and some more questions: "Have you checked/measured whether they're actually faulty?" Yes I have, in the picture I have provided on the bottom there are the ends of 3 transistor leads. This was one that fell off as I removed the card and by far the worst corroded, so despite the fact that the rest still work. I need to replace this one. " I would suspect that these transistors are pre-drivers for the power transistors and may not be super critical to get an exact replacement." Would an NPN or PNP with the same HFE suffice, or do I need to know more? Forgive me but I am a complete idiot when it comes to this kind of circuit design and their tolerances. "Did you get manuals with the printer? If so it should have ALDs (schematics)."Unfortunately not, and nor can I find any online. "The 3286 printers and 3277 terminals where infamous for intermittent problems"Excellent, glad I bought something that will provide the true IBM experience. Thank you, Al
Re: IBM transistor replacements
On Fri, 2021-11-26 at 22:33 -0800, Brent Hilpert via cctalk wrote: > On 2021-Nov-26, at 1:52 PM, Al via cctalk wrote: > > > A while ago I received an IBM 3286 printer, annoyingly some of the > > transistors in the printer section have been corroded. What I am > > having trouble with is reading the part codes and finding a modern > > equivalent of them. > > There are two types. > > > > One has a Ti logo and two sets of numbers (attached). Does anybody > > know which numbers are the part numbers and if they are IBM house > > numbered? > > Photos of the Ti transistors and card assembly: > > https://drive.google.com/drive/folders/1JTnsWS4A8NrYJNFsqK2ejOeqqJNjoF1w?usp=sharing > > > > Have you checked/measured whether they're actually faulty? Yes, they > look terrible, but that doesn't mean they're faulty. It looks like > minor corrosion of the steel case under its plating; unless it's all > the way through the die is probably fine and cozy inside. The > electrically active parts of the transistor typically aren't as > susceptible to corrosion as the steel case is. > > (Notable exception to this being 70's-era TI ICs with plated steel > pins. I've also seen some Motorola ICs with corroded pins). The Computer History Museum in Sunnyvale, CA has two IBM 1401's that are in operating order. One of the problems they had in getting one of them to work was that some of the IBM transistors also have plated steel leads that had rusted through. Ask Rob Garner , the leader of the 1401 restoration project, if he recognizes the numbers.
Re: IBM transistor replacements
On 2021-Nov-26, at 1:52 PM, Al via cctalk wrote: > A while ago I received an IBM 3286 printer, annoyingly some of the > transistors in the printer section have been corroded. What I am having > trouble with is reading the part codes and finding a modern equivalent of > them. > There are two types. > > One has a Ti logo and two sets of numbers (attached). Does anybody know which > numbers are the part numbers and if they are IBM house numbered? > Photos of the Ti transistors and card assembly: > https://drive.google.com/drive/folders/1JTnsWS4A8NrYJNFsqK2ejOeqqJNjoF1w?usp=sharing Have you checked/measured whether they're actually faulty? Yes, they look terrible, but that doesn't mean they're faulty. It looks like minor corrosion of the steel case under its plating; unless it's all the way through the die is probably fine and cozy inside. The electrically active parts of the transistor typically aren't as susceptible to corrosion as the steel case is. (Notable exception to this being 70's-era TI ICs with plated steel pins. I've also seen some Motorola ICs with corroded pins).
Re: IBM transistor replacements
On 2021-11-26 9:11 p.m., Jon Elson via cctalk wrote: On 11/26/21 3:52 PM, Al via cctalk wrote: Dear all, A while ago I received an IBM 3286 printer, annoyingly some of the transistors in the printer section have been corroded. What I am having trouble with is reading the part codes and finding a modern equivalent of them. There are two types. One has a Ti logo and two sets of numbers (attached). Does anybody know which numbers are the part numbers and if they are IBM house numbered? Photos of the Ti transistors and card assembly: https://drive.google.com/drive/folders/1JTnsWS4A8NrYJNFsqK2ejOeqqJNjoF1w?usp=sharing The other is an IBM transistor with what appears to be a 3 digit part code; for which I have been pointed to a list of house numbering and equivalents on the 1401 website. I was told the list is much older and I fear it might be out of date. However there is an entry for the part codes on my IBM transistors. Should I trust the list for replacements? http://ibm-1401.info/BobEricksonIBM-Transistor-Substitution.html The cards appear to be SLT-style fabrication, it seems there may even be SLT modules on one of them (the 1/2" square aluminum-capped blocks.) The 1401 list would be for SMS cards, they were yellow paper-phenolic cards with TO-5 and TO-18 discrete Germanium transistors. The SLT vintage were generally all Silicon transistors. Almost all this stuff had IBM house numbers. Jon The 3286 printer was part of the first generation 3270 system with 3271 & 3272 controllers and 3277 display stations, so this is early 1970s technology the packaging is similar to SLT but the chip technology was not like the hybrid SLT modules as used in 360 systems. I would agree that the parts would be all house numbered. My guess is that the 139 is the type number and the other number is a date code it seems to me that this board is the drivers for the stepper motor that moves the carriage, but it has been nearly 40 years since I worked on one of these printers. I would suspect that these transistors are pre-drivers for the power transistors and may not be super critical to get an exact replacement. Did you get manuals with the printer? If so it should have ALDs (schematics). One thing I remember about the 3284 (similar but slower) and 3286 is print timing was derived from two emitter wheels on the extreme left and if you got them to far out of adjustment, the carriage would take off and slam into the right frame and then come back and slam into the left frame making enough noise to get the attention of the customer. These printers where replaced by the 3287 which was much easier to maintain, and while it was designed for the "New Display system" they where available with a interface card that would allow them to work with 3271 and 3272 controllers. The 3286 printers and 3277 terminals where infamous for intermittent problems that you would try to isolate by swapping some of the cards with another terminal or printer and leave a note in both explaining what cards where swapped and why. It was not uncommon to old notes half buried in dust in the bottom of 3277 terminals that went back years. Paul.
Re: IBM transistor replacements
On 11/26/21 3:52 PM, Al via cctalk wrote: Dear all, A while ago I received an IBM 3286 printer, annoyingly some of the transistors in the printer section have been corroded. What I am having trouble with is reading the part codes and finding a modern equivalent of them. There are two types. One has a Ti logo and two sets of numbers (attached). Does anybody know which numbers are the part numbers and if they are IBM house numbered? Photos of the Ti transistors and card assembly: https://drive.google.com/drive/folders/1JTnsWS4A8NrYJNFsqK2ejOeqqJNjoF1w?usp=sharing The other is an IBM transistor with what appears to be a 3 digit part code; for which I have been pointed to a list of house numbering and equivalents on the 1401 website. I was told the list is much older and I fear it might be out of date. However there is an entry for the part codes on my IBM transistors. Should I trust the list for replacements? http://ibm-1401.info/BobEricksonIBM-Transistor-Substitution.html The cards appear to be SLT-style fabrication, it seems there may even be SLT modules on one of them (the 1/2" square aluminum-capped blocks.) The 1401 list would be for SMS cards, they were yellow paper-phenolic cards with TO-5 and TO-18 discrete Germanium transistors. The SLT vintage were generally all Silicon transistors. Almost all this stuff had IBM house numbers. Jon
