Date
1 - 20 of 26
OT: Analog Computer Comdyna GP-6
|
Wolfgang Schraml
Hello,
I recently took an interest in analog computers. I ended up buying two Comdyna GP-6 units, serial numbers 269 and 019. For #269, the available documentation (I spent hours googling and downloading PDFs) is pretty close to the physical unit. The one I have seems to be a little bit of a Frankenstein device since it has various boards from what looks originated out of very different production times. It also shows signs of heavy use, repairs, and potentially modifications at some point in its life. But again, the two versions of manuals I have applied to it very closely when read in combination. I am at a point where I think most problems with it (broken/cold solder joints, bad Op-Amp in the time base, missing banana jacks, broken knobs) have been or will soon be resolved, and all that's left is a thorough calibration (still trying to decipher the manuals in that regard). Now to the second unit, serial #019. It is a very different design from the other one; none of the boards are the same or look even close. The built-in voltmeter is a nixie type (working), and even the capabilities are much less than serial #269. For example, the whole block of push buttons that select "IC", HD", "OP", and "RO" is not there yet. So as the serial number would indicate, it is a very early unit. None of the documentation I found is even close. I contacted a YouTuber who has a video about one of these, and he also was not successful in obtaining any early Comdyna documentation. So my question is - does anyone have any documentation for these very early units? Or any other Comdyna related info that would give me a starting point for a restoration? If you have anything in paper, I would gladly offer to scan it in and make it available for others to use. Thank you, Wolfgang Schraml, KI7PFX PS: There is a little bit of a connection to the Tek forum … I ended up buying a Tektronix 455 to get clean outputs in XY-mode compared to my Rigol 1054Z). And a second one is on the way as a spare/backup unit.
|
|
|
|
Ed Breya
I also caught the analog computer bug a few years ago, and studied the Comdyna stuff and lots of others from the old days. I started a project to build a sophisticated analog computer with a wide variety of functions, from scratch, with more "modern" components. I have it partly built, but it's a long way to go, so it may never be finished. Instead of lots of banana jacks, the plan is to use a number of those plastic plug in IC etc prototyping plates, with the function circuits wired to the bottom, inside, and programming with wire jumpers on top. It's very compact, and provides hundreds of connection nodes.
Speaking of X-Y displays, there happens to be a somewhat related thread, regarding possible uses for an old 528 TV waveform monitor. I suggested making it an X-Y monitor, here: https://groups.io/g/TekScopes/message/173941 You should view the whole thread for context. Ed
|
|
|
|
stevenhorii
I have what I think is an unusual Comdyna. It's a hybrid analog-digital
system. Half of it looks like most other Comdyna computers I have seen, but there's another part on the bottom (the whole thing sits in an upright rack of sorts) that has digital processing in it. I think it took the analog output and digitized it, but it seems to do more than just A-D conversion. It needs work and is on my "to-do" list, but I have zero documentation on it. When I found this, the seller had two so I called a friend of mine and he wanted one so I bought both and had one shipped out to him. He got his working - I think his worked from initial power up. He's got several analog computers - he also has the large Moog and Buchla analog synthesizers. He has used the analog computers both for computing (he showed me his implementation of the logistic difference equation on one of the analog computers and the plot of the transition to chaotic behavior) and for generating music. He has an EAI TR10 and pretty much rebuilt it - capacitors were bad as were some of the transistors. He told me that for the older analog computers (and synthesizers for that matter) that such rebuilding would likely be necessary. Sounds like old Tek scopes. Steve Horii On Sun, Nov 29, 2020 at 1:37 PM Ed Breya via groups.io <edbreya= yahoo.com@groups.io> wrote: I also caught the analog computer bug a few years ago, and studied the
|
|
|
|
tekscopegroup@...
Electronic Experimenters Handbook Winter 1974:
Build the Bouncing Ball Analog Computer https://miwww.com/eeh/index.html
|
|
|
|
greenboxmaven
The subject of analog computers has brought a few thoughts forward. I remember them being offered in the classified ad sections of many magazines decades ago, and Heathkit offered a notable one. General Electric offered one in their basic experimenter's kits they sold in the early 1960s. Did anyone question the name of "computer" for these devices, instead of calculator, or analyzer? One definition I have heard and embrace is "a computer is a control or calculating apperatus with a stored operating program". I'm not trying to start a semantics dispute, but do wonder if the word "computer" has been applied at times for it's "Gee! Whiz!- It's a thinking machine! " meaning to much of the public.
toggle quoted messageShow quoted text
Bruce Gentry KA2IVY
On 11/30/20 11:09 AM, tekscopegroup@miwww.com wrote:
Electronic Experimenters Handbook Winter 1974:
|
|
|
|
Harvey White
IIRC, analog computers predated digital computers, and may well have been simply (and firstly) called computers (with anything else called a calculator).
toggle quoted messageShow quoted text
People redefine terms based on what they know at the time, often ignoring history. How many "analog" computers are on the market currently? Hence, we "define" a computer as a device with a stored program. Everyone forgets the "digital" part. Look at the missile tracking and gun aiming aparatus from WWII, analog servos, differential synchros, etc. All performed mathematical (and often quite complex) functions involving differential equations. Hmmm.... takes a digital computer for that....... And no, to answer the question, one typically qualified the word "computer" with the word "analog" if it were one, back in the days when analog computers roamed the earth. See if you can find the SF novel "Venus Equilateral" by George O. Smith. Pay attention to the electronics and computing part if you would, which shows (amongst other things) mechanical tracking "computers" using precisely machined and lapped cams. No digital involved, IIRC. Harvey
On 11/30/2020 12:05 PM, greenboxmaven via groups.io wrote:
The subject of analog computers has brought a few thoughts forward. I remember them being offered in the classified ad sections of many magazines decades ago, and Heathkit offered a notable one. General Electric offered one in their basic experimenter's kits they sold in the early 1960s. Did anyone question the name of "computer" for these devices, instead of calculator, or analyzer? One definition I have heard and embrace is "a computer is a control or calculating apperatus with a stored operating program". I'm not trying to start a semantics dispute, but do wonder if the word "computer" has been applied at times for it's "Gee! Whiz!- It's a thinking machine! " meaning to much of the public.
|
|
|
|
Jack Reynolds
Several years ago my wife and I were returning home from Florida through the panhandle. We saw signs for the USS Alabama which is anchored in Mobile Bay and decided to detour and visit the ship. i heartily recommend for anyone with an interest in engineering or computers to tour any WWII battleship they can and pay particular attention to the "Gun Director" rooms which were located behind each of the main batteries. While the main use for battleships in WWII was to place massive firepower on land targets to facilitate invasion, that is not what they were designed for! They were designed to hit other battleships over the horizon and hence invisible while both ships were moving. In addition to the obvious parameters which had to be considered, atmospheric conditions like temperature, barometric pressure and wind direction were also important. The firing solutions were all determined and implemented through the use of mechanical computers and servos. Once a firing solution had been generated, the ship could alter course and speed to become a more difficult target and the guns would still hit their mark! While so much of today's computer programming consists of using commands and syntax in a language or operating system already in existence, just imagine what it would take to solve this problem with modern stuff and then considere doing so with an electromechanical system designed from scratch in the 1930's. No Tektronix oscilloscopes were used on those ships or were injured in preparing this text!
toggle quoted messageShow quoted text
Jack Reynolds
On 11/30/2020 12:47 PM, Harvey White wrote:
IIRC, analog computers predated digital computers, and may well have been simply (and firstly) called computers (with anything else called a calculator).
|
|
|
|
Jean-Paul
From the past.....City College of New York, Terman Engineering bldg ...1964..1971:
Digital computer was an IBM 360, a huge room in the basement. My last attempt at software keypunching Fortran II...printer as big as a Volkswagen and about as noisy... The analog computing lab, had EAI with dozens (hundreds?) of Philbrick twin 12AX7 op amps and many meters and plotters. Huge patch panels to program. Occupied the entire lab wall, required a dedicated air conditioner. There was also an S plane plotter using electro graphic paper to make Bode or S plane plots. I am an internet dinosaur.... Jon
|
|
|
|
stevenhorii
Well, there were people called "computers". If you've seen the film "Hidden
Figures" the women who did the various calculations of orbital parameters and trajectories were referred to as "computers" though a little-used term for people who did computations was a "computor". No doubt that humans have "stored programs". You could argue that the patchboard used in analog computers was a "program" but analog computers did not have Boolean logic (so far as I know - maybe some of the digital/analog hybrids allowed for that). If anyone knows of an analog computer that had a way of programming an "IF...THEN" type of instruction, I would like to know about it. I think some of the analog music synthesizers could sort of do this - a voltage-controlled oscillator that could change the input and output of other analog processing stages. I have also seen some analog computer modules (I don't recall the manufacturer, but they were usually built around a single vacuum tube) with unusual names, my favorite being a "Chaostron". I have no idea what it did. Steve H. On Mon, Nov 30, 2020 at 12:47 PM Harvey White <madyn@dragonworks.info> wrote: IIRC, analog computers predated digital computers, and may well have
|
|
|
|
Ed Breya
Steve, the digital circuitry in the bigger Comdyna box could be there for a number of reasons. By the end of the analog computer era, digital stuff - computing and otherwise - had already been around for a long time, so it was natural to combine them and get the synergy of both technologies. Lacking any documentation, you may be able to at least tell kind of what the extra stuff was for, by studying the types of parts included, and doing a little reverse engineering to see what things go back and forth between sections. Here are a few thoughts on what it could be:
1. Digital readout of values. In the old days, it took quite a lot of circuitry to build a DVM - even a whole board full of stuff, for say a four digit A-D converter, and more stuff to select which signals to look at, etc. The digital results of old-time DVMs were often provided to external equipment such as digital computers for further processing, storage, and printout. These were parallel interfaces, with four bits per digit, plus possibly range and polarity info, and handshake lines. Also, sometimes the computer could control the DVM's ranges. If there's quite a large (pin count) interface connector/port on there somewhere, this could be a possibility. 2. Digital storage/equivalent time conversion for display. One of the same reasons as today. Digitizing the X-Y or T-Y signals could save and display the results of a very slow compute cycle, or present flicker-free display of a medium speed repetitive cycle. 3. Function generator. This is not the kind typically pictured when we hear the name today, that generates CW square, triangle, and sine waves. In the olden days there were all sorts of schemes to include multiplying and dividing, and nonlinear and transcendental functions in analog computers, from mechanical servos and cams, to semiconductor device characteristics, to stepped diode (even tube diodes!) line/curve fit approximations. Then there were special analog hybrids and ICs that could do log amps, and more complicated combinations for all sorts of functions. All this can be done discretely, but ICs greatly simplified it. With the availability of more digital stuff like DACs, ADCs, and ROMs, it became possible to make an arbitrary function generator. A tracking ADC follows an analog input signal, and its result addresses a ROM, which is programmed for various functions. The ROM data out drives the DAC, resulting in the desired analog function response output. I vaguely recall seeing info about this kind of FG for analog computers, but not necessarily related to Comdyna. I think by this time, the digital computers were common enough and capable enough to leave analog ones in the dust. The look-up table method is of course commonly used today for AWGs and such. I think that every modern DDS chip has buried somewhere inside, a quarter-sine ROM table and DAC to produce the desired sine output. If the Comdyna box has no microprocessor, but one or more ROMs, and ADCs or DACs, then it could be a digital function generator. 4. The last category is one where a digital computer was used to set up an analog computer - anything from the overall layout, to initial conditions, to timing control. The analog computer would then do its thing, effectively and relatively quickly, then the results would be digitized and read back to the digital computer. This seems crazy today, but there was a time when early digital just didn't have the compute power or memory, and an analog one could be used as a peripheral device to make it easier - like a math co-processor or DSP chip. If the Comdyna box has lots of relays or JFETs (set up as signal switches), or lots of more modern IC analog switches, then it could be for remote control/peripheral use, or pre-programed setups. I do vaguely recall some Comdyna literature about something like this. Ed
|
|
|
|
stevenhorii
Ed,
Thanks for the suggestions! I will dig a bit further into the box when I can get to it - it's sitting behind several large boxes of stuff. Steve H. On Mon, Nov 30, 2020 at 3:42 PM Ed Breya via groups.io <edbreya= yahoo.com@groups.io> wrote: Steve, the digital circuitry in the bigger Comdyna box could be there for
|
|
|
|
toby@...
On 2020-11-30 3:15 p.m., stevenhorii wrote:
Well, there were people called "computers". If you've seen the film "HiddenI like the definition Bruce cited because it would include such things. but analog computers did not have Boolean logic (so far as I knowComposition of that sort is the entire basis of analog computing. Analog comparators and switches (and multiplexers) have been known since valve days, and are documented analog computer functions. --Toby I have also seen some analog computer modules (I don't recall the manufacturer, but they were usually built around a single vacuum
|
|
|
|
toby@...
On 2020-11-30 3:41 p.m., Ed Breya via groups.io wrote:
...The EIA stuff used piecewise diodes and had programmable DFGs. The manual includes the complicated procedure of reducing an arbitrary function to coefficients within the component constraints. See page 32+ of http://www.bitsavers.org/pdf/eai/tr-20/EAI_TR-20_Computer_User_Handbook_Jun67.pdf --Toby
|
|
|
|
Dan Fish
I was in the Navy in the 60s as a FT (gun fire controlman). The analog computer that aimed the guns was the Mark 1 Able (MK1A). While not as accurate as digital methods, analog computers have the advantage of continuous output solution rather than recomputing solutions at some rate. See Wikipedia:
https://en.wikipedia.org/wiki/Mark_I_Fire_Control_Computer Dan
|
|
|
|
Jim Ford
And, come to think of it, my parents' 1971 Volvo had a hybrid or analog computer under the driver's seat. It was one of the first cars with fuel injection, and in retrospect, that may have had to do with its function.Jim Ford Sent from my Verizon, Samsung Galaxy smartphone
toggle quoted messageShow quoted text
-------- Original message --------From: "Ed Breya via groups.io" <edbreya=yahoo.com@groups.io> Date: 11/30/20 12:41 PM (GMT-08:00) To: TekScopes@groups.io Subject: Re: [TekScopes] OT: Analog Computer Comdyna GP-6 Steve, the digital circuitry in the bigger Comdyna box could be there for a number of reasons. By the end of the analog computer era, digital stuff - computing and otherwise - had already been around for a long time, so it was natural to combine them and get the synergy of both technologies. Lacking any documentation, you may be able to at least tell kind of what the extra stuff was for, by studying the types of parts included, and doing a little reverse engineering to see what things go back and forth between sections. Here are a few thoughts on what it could be:1. Digital readout of values. In the old days, it took quite a lot of circuitry to build a DVM - even a whole board full of stuff, for say a four digit A-D converter, and more stuff to select which signals to look at, etc. The digital results of old-time DVMs were often provided to external equipment such as digital computers for further processing, storage, and printout. These were parallel interfaces, with four bits per digit, plus possibly range and polarity info, and handshake lines. Also, sometimes the computer could control the DVM's ranges. If there's quite a large (pin count) interface connector/port on there somewhere, this could be a possibility.2. Digital storage/equivalent time conversion for display. One of the same reasons as today. Digitizing the X-Y or T-Y signals could save and display the results of a very slow compute cycle, or present flicker-free display of a medium speed repetitive cycle.3. Function generator. This is not the kind typically pictured when we hear the name today, that generates CW square, triangle, and sine waves. In the olden days there were all sorts of schemes to include multiplying and dividing, and nonlinear and transcendental functions in analog computers, from mechanical servos and cams, to semiconductor device characteristics, to stepped diode (even tube diodes!) line/curve fit approximations. Then there were special analog hybrids and ICs that could do log amps, and more complicated combinations for all sorts of functions. All this can be done discretely, but ICs greatly simplified it. With the availability of more digital stuff like DACs, ADCs, and ROMs, it became possible to make an arbitrary function generator. A tracking ADC follows an analog input signal, and its result addresses a ROM, which is programmed for various functions. The ROM data out drives the DAC, resulting in the desired analog function response output. I vaguely recall seeing info about this kind of FG for analog computers, but not necessarily related to Comdyna. I think by this time, the digital computers were common enough and capable enough to leave analog ones in the dust. The look-up table method is of course commonly used today for AWGs and such. I think that every modern DDS chip has buried somewhere inside, a quarter-sine ROM table and DAC to produce the desired sine output.If the Comdyna box has no microprocessor, but one or more ROMs, and ADCs or DACs, then it could be a digital function generator.4. The last category is one where a digital computer was used to set up an analog computer - anything from the overall layout, to initial conditions, to timing control. The analog computer would then do its thing, effectively and relatively quickly, then the results would be digitized and read back to the digital computer. This seems crazy today, but there was a time when early digital just didn't have the compute power or memory, and an analog one could be used as a peripheral device to make it easier - like a math co-processor or DSP chip.If the Comdyna box has lots of relays or JFETs (set up as signal switches), or lots of more modern IC analog switches, then it could be for remote control/peripheral use, or pre-programed setups. I do vaguely recall some Comdyna literature about something like this.Ed
|
|
|
|
Wolfgang Schraml
Hi Steve,
toggle quoted messageShow quoted text
You may find clues about what your Comdyna system is here: http://www.dvq.com/oldcomp/analog/comdyna.htm Lots of pictures of various Comdyna computers and accessories. Good luck, Wolfgang --------------------------------------------------------
On Sun, Nov 29, 2020 at 12:39 PM, stevenhorii wrote:
|
|
|
|
Roy Thistle
On Sun, Nov 29, 2020 at 10:36 AM, Ed Breya wrote:
by that phrase, you mean a "breadboard?"... which is certainly not a board, nor has anything much to do with bread; but, is as descriptive as calling a peanut a nut. There might be a reason that programmable analog computer manufacturers chose that classic patch panel design (you see that in the Comdyna units). To whit... for starters...does an insertion point, on a breadboard, adequately model a relatively isolated plug and socket? (the interconnect we know meets the specification.)
|
|
|
|
Ed Breya
Yes Roy, I think "breadboard" or "solderless breadboard" are the names commonly applied to those things. When I was writing the message, my brain was in deep memory fetch mode, and the name escaped me.
Regarding patch panel layout and connectors, the early analog computers used what was commonly available, like banana jacks and plugs, and the machines were vacuum tube based, and quite large, so space was not so much an issue. Also, the signals ranged to 100 V or more (a railed amplifier might reach 200-300V), so the user could literally be shocked by the computation results. So, nice big jacks and insulated jumpers made sense, and on a spacious panel there's some room for labeling the nodes and functions. I think this became the tradition for all that followed, even after the evolution to solid-state and lower, safer voltage levels. I think EAI and maybe others came up with smaller types of plugs and jacks in later models, making it more compact, but the downside is less room for labeling. Mine has high density for lots of nodes, and no plugs and jacks are needed - just little jumper wires with stripped ends. But, there's no room for labels. I'm using the kind of breadboards that have little index numbers and letters marked right on them, and will have to make an index sheet that shows what things are where. There are pluses and minuses with any approach. Ed
|
|
|
|
stevenhorii
Wolfgang,
toggle quoted messageShow quoted text
Wow! Thanks - I believe after looking at the pictures on DVQ that what I have is an LGP-20. If I am not mistaken, the Pomona Electronics dual banana plug parts will exactly fit the spacing the jacks, at least on the Comdyna computers. My friend who uses one of the Comdyna GP-6 computers looks for Pomona banana plug stuff all the time. DVQ also has the Technical Measurements Corporation "CAT-400". I believe that to be one of the first digital oscilloscopes if not the first. It was designed to average signals to improve S/N and has core memory in it. it is all transistorized. It was not a general-purpose scope by any means. These were mostly used by electrophysiology folks who were studying single-nerve or small nerve bundle impluses. They would put a pre-amp before the CAT-400 and then use the CAT to digitize and average signals to get better S/N. The company also made a multichannel analyzer and an autocorrelator, both in the same-sized (and same color scheme) cases. They were quite expensive - about USD $20,000 in the mid 1960s. Steve Horii
On Mon, Nov 30, 2020 at 9:06 PM Wolfgang Schraml <wschraml@gmail.com> wrote:
Hi Steve,
|
|
|
|
Jim Ford
I don't know much about analog computers or electrophysiology, but I do know that the standard spacing of banana jacks and plugs is 3/4 inch. HTH.Jim Ford Sent from my Verizon, Samsung Galaxy smartphone
toggle quoted messageShow quoted text
-------- Original message --------From: stevenhorii <sonodocsch@gmail.com> Date: 12/1/20 1:12 PM (GMT-08:00) To: TekScopes@groups.io Subject: Re: [TekScopes] OT: Analog Computer Comdyna GP-6 Wolfgang,Wow! Thanks - I believe after looking at the pictures on DVQ that what Ihave is an LGP-20. If I am not mistaken, the Pomona Electronics dual bananaplug parts will exactly fit the spacing the jacks, at least on the Comdynacomputers. My friend who uses one of the Comdyna GP-6 computers looks forPomona banana plug stuff all the time.DVQ also has the Technical Measurements Corporation "CAT-400". I believethat to be one of the first digital oscilloscopes if not the first. It wasdesigned to average signals to improve S/N and has core memory in it. it isall transistorized. It was not a general-purpose scope by any means. Thesewere mostly used by electrophysiology folks who were studying single-nerveor small nerve bundle impluses. They would put a pre-amp before the CAT-400and then use the CAT to digitize and average signals to get better S/N. Thecompany also made a multichannel analyzer and an autocorrelator, both inthe same-sized (and same color scheme) cases. They were quite expensive -about USD $20,000 in the mid 1960s.Steve HoriiOn Mon, Nov 30, 2020 at 9:06 PM Wolfgang Schraml <wschraml@gmail.com> wrote:> Hi Steve,>> You may find clues about what your Comdyna system is here:> http://www.dvq.com/oldcomp/analog/comdyna.htm>; Lots of pictures of various Comdyna computers and accessories.>> Good luck, Wolfgang>> -------------------------------------------------------->> On Sun, Nov 29, 2020 at 12:39 PM, stevenhorii wrote:>> >> > I have what I think is an unusual Comdyna. It's a hybrid analog-digital> > system. Half of it looks like most other Comdyna computers I have seen,> but> > there's another part on the bottom (the whole thing sits in an upright> rack> > of sorts) that has digital processing in it. I think it took the analog> > output and digitized it, but it seems to do more than just A-D> conversion.> > It needs work and is on my "to-do" list, but I have zero documentation on> > it. When I found this, the seller had two so I called a friend of mine> and> > he wanted one so I bought both and had one shipped out to him. He got his> > working - I think his worked from initial power up. He's got several> analog> > computers - he also has the large Moog and Buchla analog synthesizers. He> > has used the analog computers both for computing (he showed me his> > implementation of the logistic difference equation on one of the analog> > computers and the plot of the transition to chaotic behavior) and for> > generating music. He has an EAI TR10 and pretty much rebuilt it -> > capacitors were bad as were some of the transistors. He told me that for> > the older analog computers (and synthesizers for that matter) that such> > rebuilding would likely be necessary. Sounds like old Tek scopes.> >> > Steve Horii> >> > On Sun, Nov 29, 2020 at 1:37 PM Ed Breya via groups.io <edbreya=> > yahoo.com@groups.io> wrote:> >> > > I also caught the analog computer bug a few years ago, and studied the> > > Comdyna stuff and lots of others from the old days. I started a> project to> > > build a sophisticated analog computer with a wide variety of functions,> > > from scratch, with more "modern" components. I have it partly built,> but> > > it's a long way to go, so it may never be finished. Instead of lots of> > > banana jacks, the plan is to use a number of those plastic plug in IC> etc> > > prototyping plates, with the function circuits wired to the bottom,> inside,> > > and programming with wire jumpers on top. It's very compact, and> provides> > > hundreds of connection nodes.> > >> > > Speaking of X-Y displays, there happens to be a somewhat related> thread,> > > regarding possible uses for an old 528 TV waveform monitor. I suggested> > > making it an X-Y monitor, here:> > >> > > https://groups.io/g/TekScopes/message/173941>; > >> > > You should view the whole thread for context.> > >> > > Ed> > >> > >> > >> > >> > >> > >> >>>> >>>
|
|
|