Scope simulation

[Angelo Papenhoff/aap]

This thing here looks absolutely gorgeous and it would be really cool to use with a PDP-1: https://oscilloscopemusic.com/software/oscilloscope/

perhaps it can be modified to have a nice P7 phosphor and hooked up to the type 30 simulation? :)

[Unibus]

Hi,

At one stage I had various early Tektronix scopes and P7 phosphor was an option on different models over the years. Your chances of finding one these days are slim to non-existent. Might be worthwhile asking around, you might get lucky.

Regards,

Garry

On Thu, 12 Mar 2026 at 21:18, Angelo Papenhoff/aap <a...@papnet.eu> wrote:

This thing here looks absolutely gorgeous and it would be really cool to use with a PDP-1: https://oscilloscopemusic.com/software/oscilloscope/

perhaps it can be modified to have a nice P7 phosphor and hooked up to the type 30 simulation? :)

--
You received this message because you are subscribed to the Google Groups "[PiDP-1]" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pidp-1+un...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/pidp-1/4c70d37d-4a06-4020-a984-7e67a032b4d5n%40googlegroups.com.

[Bill E]

I'm feeling a slight urge to do some hardware design. I've been pondering a little board that would provide 2 10-bit dacs to drive a real scope and a real lightpen, probably usb connected, but maybe a tcp connection would work. The goal would be an interface that could get lightpen hit coordinates back in time to really be guaranteed to be in sync with the dpy that was executed. I think the current mouse imitation does manage that, but I haven't bothered to find out. My implementation and Angelo's are almost identical, so apparently we both thought what we have now is the best we could do.

Not sure I'll get around to it, and not sure how useful it would be. I doubt people are going to want to put a scope along with their -1, although some of the new small ones with an LCD panel might be ok.

BTW, does anyone remember the old Vectrex game console? That was an impressive vector-drawing graphics platform driven by a 6502 processor. Maybe hack one of those for a display? And yes, I have one, like a bunch of other obsolete junk, er, history.

Bill

[Oscar Vermeulen]

That would be nice! The CHM PDP-1 guys want one, it's a request of theirs in fact.

[Unibus]

Hi Bill,

I was hoping to use the I2C or the 3.3V serial connectors that are unused after completing the Console build. Then route IOT instructions to communicate with external hardware i.e. not using WiFi or Ethernet.

Silicon Chip magazine has had a versatile SBC based on the Raspberry Pi 2350B that can be ordered directly from the Chinese manufacturer. See:-

https://www.siliconchip.com.au/Issue/SC/2025/November/RP2350B+Computer

This SBC supports Analog and Digital operations plus it provides HDMI, mouse and keyboard control as a standalone PC. Catches are this SBC does not have Ethernet, plus USB programming is probably too painful for me.

Does this level of hardware interest you?

Regards,

Garry

 

--

You received this message because you are subscribed to the Google Groups "[PiDP-1]" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pidp-1+un...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/pidp-1/f9eeddb7-00ca-4c64-8c3a-4d8336cd5790n%40googlegroups.com.

[Lawrence Harris]

I built some things like this back in my university days to drive a scope from a Z80 using 8 bit DACs.  Didn’t do a light pen but that aught to be fairly straightforward.  What I didn't’ do is add the op amps and slew control so that when I draw a vector between 2 points I get a straight line.  I had to plot all the points along the line to get something decent.  Last week I was trying to duplicate it with an ESP32S3 development board - I think it’s too fast, I’ll need to delay the point output so I can see them on the scope before the next point is drawn. :-)

ps.  I have not looked at the PDP-1 display to know if it did full vector operations.  I am assuming it did.

pps.  The Vectrix service manual with a parts list and schematics is at : https://console5.com/techwiki/images/a/a7/Vectrex-Service_Manual.pdf

Lawrence

--
You received this message because you are subscribed to the Google Groups "[PiDP-1]" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pidp-1+un...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/pidp-1/f9eeddb7-00ca-4c64-8c3a-4d8336cd5790n%40googlegroups.com.

[Adam Thornton]

The Vectrex is a 6809.  I have one too but in my last attempt to fix it (the internal amp died) I have shorted something and now it eats fuses and I haven't traced the problem yet.

The Vectrex is a super cool machine, though.  It's only like 3 years later than the Atari 2600, but the sophistication of the software environment is leaps and bounds ahead.  In the 2600 you're continually racing the raster beam across the screen, and very little of your time is spent actually doing game logic (you get the little bit of horizontal refresh and the larger chunk of vertical refresh, but mostly you're stuffing values into the display registers).

But on the Vectrex...you load vectors into shape tables, you set up music pitches in a region of memory, and then you have a luxurious 50,000 cycles between every frame to do whatever you want with, and the firmware manages reading those shapes out, scaling them like you told it to, and putting them on the screen.  It's quite lovely.

Adam

--

You received this message because you are subscribed to the Google Groups "[PiDP-1]" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pidp-1+un...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/pidp-1/f9eeddb7-00ca-4c64-8c3a-4d8336cd5790n%40googlegroups.com.

[Matthias Barthel]

Garry,

On the First days when i build my pidp1, i connect some gpios over the i2c connectors; this is not complicated but the i2c bus is to slow for out purposes.

Thats the reason why i developed my backplane over tcp for connecting external hardware to my pdp1's.

https://github.com/maklumatpemankanan/pidp1-backplane-mods

I have also build a pdp1 around a esp32 controller, but its not opimized yet. The backplane works on both versions. It also exists a adapter to drive the pidp1 with a esp32. I have a new version here with build it and working 16-channel sequence break system.

Best matthias

To view this discussion visit https://groups.google.com/d/msgid/pidp-1/CAP2nic2G1HAjgjvgw5KAGoLx-L4QrYkzMnWFh_MOyzErWbnrYA%40mail.gmail.com.

[Unibus]

Hi,

My problem is I've standardized on hardwired interfaces and the I/O hardware doesn't support UTP. The stupid thing is they had a much earlier version that had some form of Ethernet support. I don't have time to experiment but one thing I was going to try was a USB to Ethernet adapter then see if anything could be done with it.

Regards,

Garry

To view this discussion visit https://groups.google.com/d/msgid/pidp-1/CAHvYjvHeV_a-7%3D_uSF7hxut_LjDc%3Di0mRUsbidP1HJ5EwStCqA%40mail.gmail.com.

[Bill E]

Nope, the Type 30 is strictly a point-plotter. Load the 2 10-bit dacs, wait a while, turn on the beam for a while, turn it off. That's it. Not to say that wasn't impressive for the time.

DEC was apparently a bit embarrassed about this, because they came out with another display not that much later, early '60s,  that basically did everything imaginable, including vectors and DMA right from user memory, the Type 340. There seems to be a popular belief that the Type 30 was a vector display, but that's not true.

Bill

[Norbert Landsteiner]

I think, some later models of the Type 30 did support vector (or there was at least an option), but most of them didn't and certainly not the earlier ones.

(E.g., the 30E, for which we have the manual and which is contemporary to the PDP-8 (i.e. 1965 or later), is (still) strictly point plotting. But I'm not so sure about the G and H.)

Notably, intensity (z) is achieved by varying activation time, meaning, the activation is at a constant level, regardless of the intensity bits. The real complexity of the display is probably in the dampening circuitry, which prevents overshoot and undershoot of the display location, thus providing the precision ant image stability. (I guess, this contributes quite a bit to the 50 µs gap between individual plots.)

Norbert

[Ian Schofield]

Hi all,

 The beam settling time of 50us was (surprisingly) because the CRT has magnetic deflection see:  digital-precision-crt-display-type-30.pdf

 The next displays DEC came up with ... VR12, VR14 and VR17 also had magnetic deflection. This is something of a problem

 for high speed displays. For our GT40 we eventually bought an HP 17" electrostatic deflection tube and this was really cool!

 And, we turned the clock up a bit to get more displayed points.

 The main difference was that the HP's convergence was much better and it was possible to expand the display to full screen and still stay in focus.

 I say surprisingly as electrostatic CRTs have been around for a while but needed high voltages for the deflection circuits.

 Low voltage high current drivers are a bit easier to design hence the use in virtually all commercial TV sets.

 Also, it remains possible that the tube they used for the display which was used for radar may have been designed for use with a rotating yoke to get a PPI display.

 Although, most systems used a servo driven sin/cos potentiometer that rotated with the antenna to drive the deflection amplifiers.

Regards, Ian.

[Unibus]

Hi,

For the GT40 something I suspected but wouldn't be able to prove is the GT40's impact on the design of the PDP-11/05. At the University where I worked we seemed to be an early adopter of PDP-11 hardware. Physics Lab had a PDP-11/20 and the Computer Centre has various versions of PDP-11/05 and PDP-11/10 systems. From memory (no pun intended) the CPU of the oldest hardware was revision M or an earlier revision. I'm confident our oldest PDP-11/05 processor predated the Bitsaver's schematics and Engineering Set. The earliest hardware did not have the four 74181 ICs in a nice vertical column but they were spread out in an irregular layout and had a mass of jumpers modifying the board. The later hardware versions cleaned up the board layout so that the 74181 layout was the clean vertical column you see in the Bitsavers' documentation plus the mass of jumpers disappeared. Also the earlier documentation had a whole list of revisions documented on the schematic. Again from memory the list of revisions might have been over one column long. They all disappeared from the Bitsavers' schematics.

It appeared to me that a lot of design work for the earliest version of the PDP-11/05 had gone into bus control to support a multiprocessor configuration which was basically the GT40. Must admit at home in a BA-11 box I removed the core memory, installed a MUD backplane and had max memory using semiconductors but never tried for multiprocessor configuration. Unfortunately the University Admin staff stuffed me around and auctioned off our GT40 to some random person and I never sighted again.

Regarding the 74181 there is/was confusion on the year it was first released. I had tried to confirm when DEC first received the 74181 chips and started to prototype the PDP-11/05. I wondered if DEC had received pre-production versions of the 74181 or early production samples etc.  

Sorry I don’t knw.

The 11/05 used  the AMD  alu chip. I recall meeting with Jerry Sanders and he asked what the price needed to be, and I said it hasto be cheaper than an  implementation of smaller scale parts.

g

I had asked but the info seems to be lost.

Regards,

Garry

--

You received this message because you are subscribed to the Google Groups "[PiDP-1]" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pidp-1+un...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/pidp-1/fa6978db-89ee-406d-86ab-026474b6ff9en%40googlegroups.com.

[Clem Cole]

below

On Fri, Mar 13, 2026 at 10:13 PM Unibus <uni...@gmail.com> wrote:

Hi,

For the GT40 something I suspected but wouldn't be able to prove is the GT40's impact on the design of the PDP-11/05.

They were separate projects / different teams.  While I was not working for DEC at the time (I would later), I knew a few of the players involved in both of the groups and have heard/know many of the stories (FWIW: my good friend and colleague the late Jack Burness of "moonlander" fame had a cubicle across from me when we worked for "fossil"  - Roger Gourd). The Graphics-11 group built the GT40 and packaged it as the VT40 computer, using the PDP-11/10 >>system<< that central engineering developed.

The KD11-B CPU (used in both the 11/05 and 11/10) was driven more by requirements coming from the Laboratory Product Division (LDP) to reduce the entry price of lab equipment base systems. Their PDP-12 at $27K follow-on for the PDP-8 was well received technically, but getting eaten alive by systems built up on DG's $4K Nova (base price), in particular (the 12 and Nova were both introduced in 1969).  Note that the Nova was stripped and much more DIY (like the PDP-8 had started) and the 12 was sold as a complete system with lots of lab hardware in it.  The problem for LDP was that the base price of 11/20 of $12K in 1970, and you fleshed it out closer to the complete system, like they had with the 12; the system quickly got to $30K-$40K. So a $5K base of the 11/05 was much more in line with the NG Nova - althought it was marketed for OEM's.  The 11/10 was the "complete" system package. The Graphic group used the 11/10 as their base (and Jack was its first real programmer - as an aside, if you have never looked at the sources to Moonlander, it's a work of art; particularly its transcendental package using 16-bit integers, since the 11/05 did not support an FP [http://www.brouhaha.com/~eric/retrocomputing/dec/gt40/]). Also note that the GT40/VT40 was released in late 1972, while the 11/05 and 11/10 were in June.

Also, remember that "embedded" microprocessors for a lab are still a few years away [the Intel 4004 is 1971, 8008 is 1972, and 8080 is 1974]. Still, you won't see computer "systems" like the Altair until 1975, and it will take a few more years after that to achieve the same functionality as that available in the laboratory or graphics markets served by the PDP-8, 11, and 12 families.

 

...

Regarding the 74181 there is/was confusion on the year it was first released. I had tried to confirm when DEC first received the 74181 chips

DEC central engineering at them from TI when it was released in the spring of 1970.  AMD would become one of the firms that had a 74181 second-source license.  DEC used (and shipped) the 74181 on both 11/20 and the PDP-12 shortly after they were available in volume. They continued to use it through the 11/780 but also began designing with the AMD 2900 family after 1975.

 

and started to prototype the PDP-11/05. I wondered if DEC had received pre-production versions of the 74181 or early production samples etc. 

I can easily imagine that DEC purchased and shipped the AMD version as a second source for better pricing, but they would have qualified them for the KD11-B and would not have shipped pre-production parts in a revenue system.  By the late 1960's timeframe, DEC's engineering and manufacturing processes were fairly formal.   I could have believed that using less well-qualified parts >>might<< have happened earlier with something like the PDP-1, but by the time of the PDP-11 program, it just could not see it being possible.

[Paul Birkel]

When considering the market price of the original Nova keep in mind that it was a 4-bit ALU, not 16-bit, so four passes plus assembly to get anything done.  Example slice implementation (pre-74181 era):

[Unibus]

Hi,

My home PCs included the PDP-11/05 with PTS & CAPS-11 with LPS-11 and ASR-33. Eventually it was NatSemi SC/MP then SC/MP II, 6800D2Kit, etc. Somewhere there is a photo of a VAX-11/750 on seagrass matting in the master bedroom. Needless to say we were banished to a back bedroom.

Regards,

Garry 

[Milo Velimirović]

11/20 cxx, AMD 2901 clarification.

On Mar 13, 2026, at 8:59 PM, Clem Cole <cl...@ccc.com> wrote:

Regarding the 74181 there is/was confusion on the year it was first released. I had tried to confirm when DEC first received the 74181 chips 

DEC central engineering at them from TI when it was released in the spring of 1970.  AMD would become one of the firms that had a 74181 second-source license.  DEC used (and shipped) the 74181 on both 11/20 and the PDP-12 shortly after they were available in volume. They continued to use it through the 11/780 but also began designing with the AMD 2900 family after 1975.

Other second sources for the ‘181 would have been Signetics or others; DEC used a lot of parts from Signetics in the early days of the -11s. The 11/20 was one of the few (non-VLSI) -11s that did NOT use the ‘181. Its ALU was designed around the 7482 2-bit wide full adder and 74H53.

See https://www.bitsavers.org/pdf/dec/pdp11/1120/1120_SystemSchems_Feb70.pdf pages M224, K7-1 — K8-2

AMD 2900 chips, and in particular the 2901 saw use in the Floating Point processors for the 11/34 and 11/44. Surprisingly, (to me, anyway,) the 11/60’s FPP, the FP-11E used the 74S181 and not the AMD 2901, even though the 34 and 60 were contemporaneous. I don’t know of any DEC designed pdp11s that had a CPU built around an AMD2901; I can’t speak to VAXen.

 

and started to prototype the PDP-11/05. I wondered if DEC had received pre-production versions of the 74181 or early production samples etc. 

I can easily imagine that DEC purchased and shipped the AMD version as a second source for better pricing, but they would have qualified them for the KD11-B and would not have shipped pre-production parts in a revenue system.  By the late 1960's timeframe, DEC's engineering and manufacturing processes were fairly formal.   I could have believed that using less well-qualified parts >>might<< have happened earlier with something like the PDP-1, but by the time of the PDP-11 program, it just could not see it being possible.

— MV

[Clem Cole]

On Sun, Mar 15, 2026 at 7:42 PM Milo Velimirović <milovel...@gmail.com> wrote:

Other second sources for the ‘181 would have been Signetics or others; DEC used a lot of parts from Signetics in the early days of the -11s. The 11/20 was one of the few (non-VLSI) -11s that did NOT use the ‘181. Its ALU was designed around the 7482 2-bit wide full adder and 74H53.

See https://www.bitsavers.org/pdf/dec/pdp11/1120/1120_SystemSchems_Feb70.pdf pages M224, K7-1 — K8-2

Excellent.  Thank you.  I had always thought it was the first use of the 74181.  I have to ask my friends from the 11 CPU team.  As I said, when the 11/20 was developed, I was not yet there.

AMD 2900 chips, and in particular the 2901 saw use in the Floating Point processors for the 11/34 and 11/44.

Yep, and I suspect I know why - Jeff Mitchell was the primary CPU guy (he also did the Masscomp MC500 a few years later), and IIRC Aaron Boxer was the FP designer (and did both the Masscomp FP and the AP).  And in the cases of both FP units, Dick Munroe had a heavy hand/if not wrote most the microcode.  I don't know if he worked on the 44, but I know he wrote much of the microcode for Comet.†

As I understand it, Jeff, there was a bet at the time that they could reduce the 11/40 to a single hex board.  While the 11/40 was doing extremely well, the 34 was driven by cost reductions and better margins.  FWIW: My favorite part of that design is the front console that has an Intel 8008, which can do DMA to/from the unibus😱

Surprisingly, (to me, anyway,) the 11/60’s FPP, the FP-11E used the 74S181 and not the AMD 2901, even though the 34 and 60 were contemporaneous.

Not completely - different teams and targets,  it was sold internally as DEC marketing wanted a "commercial system" follow-on for the 11/40 (i.e., RSTS Cobol). They were trying to make a run at the 1976 release of the Burroughs 1800, which was a small-to-medium business system, which famously changed its microcode when different language systems were running (it's a very cool machine).  But Burroughs had successfully made a big deal out of this feature to customers.  So one of the requirements for the 60 was a writable control store (the stock 40 was microcoded, but with a board sole by 3-Rivers, designed by CMU for C.mmp, you could have an 11/40e with WCS.  I'm not sure how many firms bought the tools and tried to use it [we got them at Tektronix and Steve Glaser and I were starting to add the C support instructions (CSV/CRET) to our 11/60 that had been written for the CMU 11/40e, but when never finished it, as mostly orphaned the 11/60; when our boss found an 11/70 that had been running RSTS Cobol for a finance group at Tektronix and decided it was too small [they went back to IBM, we the system and that became the Teklabs 11/70].

As I understand it, the big screw-up for the 11/60 was the 40-class system limits of its MMU and lack of I/D.  The former limited the system to 256KB.  The minimum for RSTS was 128KB, but even with 256K with anything other than Basic, it was sluggish [and as I understand it from the folks we had the 11/70 from - which had 4M in it, it was still too slow for their needs]. As for the latter lack of I/D space, it was not required, but it helped a lot.   I do know that at some point, the 11/60 set the record (and I think in the end always held) for going from announcement to "traditional" products.

 

I don’t know of any DEC designed pdp11s that had a CPU built around an AMD2901; I can’t speak to VAXen.

KS10 - a.k.a. 2020 https://gunkies.org/wiki/KS10 

† I have a wonderful story about the trade made between the Masscomp and DEC hardware teams in the heat of the night WRT to microcode development, but I'll save that for some other time.

[Milo Velimirović]

I worked for a VAR briefly in the late ‘70s who sold mostly 11/34 and 11/34A systems running DIBOL applications. When the customers complained of poor performance, either from less than optimal coding/compiling or “too many” attached VT52s, the solution was to “put wings on a pig” and sell them an 11/60. Some of us who had other prior experience, particularly with UNIX, suggested deploying on 11/40s or 11/45s to solve the speed problem but those CPUs weren’t available to our VAR channel from DEC. We could have sold 11/70s, but that would have required a jump from RT-11/CTS300 to RSTS/CTS500.

It was a good learning experience for me. In hindsight it was just as well that I was let go in the big layoff. Not too long afterwards the owners became guests of the federal govt. — something about numbers that differed from reality on documents filed with the Small Business Administration.

On Mar 15, 2026, at 7:39 PM, Clem Cole <cl...@ccc.com> wrote:

As I understand it, the big screw-up for the 11/60 was the 40-class system limits of its MMU and lack of I/D.

That is my understanding too. I’ve also read/heard that this was deliberate to keep the /60 from nibbling away at the /70’s sales. The /60 was arguably a better implementation of the -11 ISA in a better package.

— Milo

Looking fwd to µcode stories.