Crowdfunding the PiDP-10 (and its size issues)

[Damien Boureille]

Hello,

Being new here, I understand the overwhelming cost of the injection mold is the most determining factor of the PiDP-10 replica size.

I read some people expressing in the past months a desire for a full-size replica, but this simply can't happen because of the resulting price (mold and shipping), and the fact its 80cm would be too cumbersome for most.

A 2:3 replica can bring the price down while still having a decent size, but I also understand some might also find it a bit on the small side for a mainframe class computer. Which I can also understand: a replica looking (to some) a little "toyish" or "smallish" would be acceptable for a PDP-8, but much less for the almighty DEC-10. This needs to be the flagship PiDP.

Since Oscar is taking a bet here with his own money and every millimeter counts, I came up with the idea of starting a crowdfunding with incremental goals for the injection mold, ranging from a smaller to a larger size.

Such goals could be:
- Goal 1. The usual PiDP 2:3 replica size (53cm). Seems like most users would already be very happy with it.
- Goal 2 could aim at a slightly bigger 3:4 replica (60cm).

This already makes a big difference in feeling closer to the "real thing" while still fitting on a keyboard tray, but also a big difference in price, both for the mold and shipping rates.

If a good share of the cost could be pre-funded, then maybe this could make such an option possible. I know I would love a 3:4 replica if a crowdfunding could make it happen!

- Goal 3 could aim at something even a little larger, like a 4:5 replica (~64cm) for those wanting something closer to the authentic KA-10, but this is probably where both the size and price would become overwhelming to most.

So this may be an absurd idea, but maybe such a crowdfunding could help make sure Oscar doesn't ever need to flip burgers :), bring the individual price of the kit down a bit, and maybe see if a slightly larger replica could become a possibility.

What do you think?

Damien

[Vince Mulhollon]

In spirit, I like the idea a lot.

In practice, I wonder if die design and manufacture has improved to the point of being as predictable as PCB mfgr.

In the distant past, the cost of producing a working, shipping, injection molding die was very unpredictable.  Its not just simple stuff like enough draft not to get stuck in the mold, but issues like differential shrinkage and warping and surface roughness etc.  Something long and skinny like this, if you pressure it hard enough to fill fast enough before the plastic cools, will it squirt out?  I would not personally care if my PDP-10 were warped a tenth of an inch "overall" but I'd notice a tenth of an inch gap between components.  Another comment about the biz is the bigger the part the harder it is to meet tolerances, so the risk increases drastically with size.  Even I could probably make an injection molded volume control knob, and all I know is I had a school friend a long time ago who aspired to this kind of work.

I would be very happy to hear that due to modern CAD/CAM etc, injection molds are now fully predictable and "boring" but they probably are not.

I find it amusing to see people complain about AI replacing "everything" yet I know for a fact that something as simple as making a car panel that fits, at least as of recently, still requires/required extensive human craftsmanship and hands on input.

[Lars Brinkhoff]

damien.b...@gmail.com wrote:

A 2:3 replica can bring the price down while still having a decent size, but I also understand some might also find it a bit on the small side for a mainframe class computer. Which I can also understand: a replica looking (to some) a little "toyish" or "smallish" would be acceptable for a PDP-8, but much less for the almighty DEC-10. This needs to be the flagship PiDP.

I would argue that even a full size PDP-10 operator console (i.e. what Oscar is replicating) looks somewhat puny and lost compared to the full CPU and IO, three cabinets in all.  If you want to go for an impressive display, probably trying to get the two processor indicator panels going would do a lot.  They have hundreds of lamps that are sure to dazzle anyone.  A real machine would also have several more boxes for memory, controllers, tapes, disks, etc.  This is the kind of bulk you would expect from a mainframe, but of course beyond the scope of a PiDP type replica.  I bet such a configuration even in 1:10 scale would be quite impressive.

[Mike Ross]

Strongly agree: I've advocated for 1:1 scale replica consoles from the start. 

But, it's not my project, it's not my $$$, and it's not my hard work!

So I'll say 'take my $$$!' to whatever emerges. 

Mike 

--
You received this message because you are subscribed to the Google Groups "PiDP-10" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pidp-10+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-10/988cb504-3ef7-4426-ad32-dae0ba80f3ccn%40googlegroups.com.

--

http://www.corestore.org
'No greater love hath a man than he lay down his life for his brother.
Not for millions, not for glory, not for fame.
For one person, in the dark, where no one will ever know or see.'

[Bill Ezell]

I will of course be buying one. To go along with my PDP-8/I and
PDP-11/70 replacements.

I was wondering if anyone knows what the power consumption for a
'typical' KA installation was? Must have been huge, given it was all
discrete components on those 'flip-chip' cards.

Bill

--
Bill Ezell
----------
You keep using that word. I do not think it means what you think it means.

[Damien B.]

KA-10: 4300W (3 cabinets weighting 875kg).

But this isn't as much as the KL-10, implemented in ECL gates with a master clock running at a blazing 50MHz. This one used 20 kW in 3 cabinets weighting 1200kg.

Of course you also need to account for the core memory modules: an MH10 (4*64 Kwords) sucked another 2.5 kW and a larger installation could have 6 of them.

I assume a large 1090 installation could use around 40 kW, and that's still without cooling.

-Damien

[Bill Ezell]

Thanks. Yep, forgot about the ECL. That logic ran in 'linear', i.e. not saturated mode like TTL. Huge power hog, huge cooling requirements. I'm actually somewhat amazed by the core mem consumption, but then again, mag core required significant switching current for the cores. While the current pulses were short, seem to remember it was in the amp range. Multiply that by the number of accesses/sec, adds up. I think my real PDP-11/05 itself took close to 1kw.

Historical note, in college I bought up random parts from some DEC guy who claimed he salvaged them from the DEC garbage dump to build my 11. Turns out he was stealing them from the production line. Got found out, went to jail. I also had an 8/I sourced from the same guy. But, I did have to assemble it all and do a bit of troubleshooting. The core memory in specific was a bit temperamental. Somewhat regret not keeping the 11, but really is a boat-anchor now given the great emulator we have.

Bill

On 3/3/2023 11:08 AM, Damien B. wrote:

KA-10: 4300W (3 cabinets weighting 875kg).

But this isn't as much as the KL-10, implemented in ECL gates with a master clock running at a blazing 50MHz. This one used 20 kW in 3 cabinets weighting 1200kg.

Of course you also need to account for the core memory modules: an MH10 (4*64 Kwords) sucked another 2.5 kW and a larger installation could have 6 of them.

I assume a large 1090 installation could use around 40 kW, and that's still without cooling.

-Damien

[Damien B.]

I went to Bitsavers and found some documentation:

http://www.bitsavers.org/pdf/dec/pdp10/KL10/EK-108OU-PD-002_KL10-Based_Physical_Description_Mar77.pdf

Page 55:

KL-10 processor : 19800 W

PDP-11/40 frontend processor (DN87S): 5360 W

MH10 core: 2400 W

RP04 Disk Pack Drive: 2100 W

So a minimal system with one disk drive and just 256 kW of memory already consumes around 30 kW.

Maybe I was still a little optimistic with 40 kW for a large system.

Here's the same for the KA-10, page 21, or 17 in the doc:

http://www.bitsavers.org/pdf/dec/pdp10/KA10/DEC_10-HAAB-D_SitePrep_May70.pdf

And now for a little interesting math:

- The KA-10 (DTL) has a power usage of 4300W and a performance of 0.38 MIPS = 11.3 kW / MIPS

- For the KL-10 (ECL): 19800W / 1.8 MIPS = 11 kW / MIPS

So both technologies appeared in fact to have about the same efficiency.

The TTL KI-10 was probably the most efficient of all, as well as the "Jupiter" KC-10, using LSI NAND-gates (like the VAX 11/750) with an F11 frontend CPU.

The most efficient PDP-10 ever is probably the FPGA-based TOAD-2 from XKL, which I think is still being produced today.

If emulation counts, then a low power ARM board running SIMH probably yields an ever better efficiency.

Damien

[Andy]

FWIW I am very excited about any PiDP-10 size, but 2:3 gets my vote as all my other PiDPs are that ratio and look great in my rack. 

How about some processor indicator panels as mentioned by Lars?!!

[Bill Ezell]

I agree, a size that is practical but still is clearly identifiable. While I really want to emulate a KL, the KA is the best console, has character, and at 2:3, it still looks good. I just can't imagine the effort that's been put into this, but I'm very grateful. The attention to detail is almost obsessive.

So, if a 1:1 was available, would I buy it? Yes, but not sure where I'd put it.

Bill

PS - for some reason there are a lot of ex-DEC engineers in my town. We have them over, they instantly rave about the 8 and 11 reproductions. When I tell them I can access via PuTTY, they are amazed. BTW, I've had RSX-11 up constantly for well over a year. After I fixed that threading issue in the SW.

On 3/3/2023 2:40 PM, Andy wrote:

FWIW I am very excited about any PiDP-10 size, but 2:3 gets my vote as all my other PiDPs are that ratio and look great in my rack. 

[Damien Boureille]

While I really want to emulate a KL, the KA is the best console

I assume the PiDP-11 could be fairly easily converted to a KL-10 if you really wanted one. It would take a bit of change in the software to have the KL10 simulator drive the LEDs, as well as a new color theme. The KL-10 face could even be sold as an addon for the PiDP-11, although changing the buttons would be more annoying.

I agree that the KA-10 was the right choice for the PiDP-10.

So, if a 1:1 was available, would I buy it? Yes, but not sure where I'd put it.

1:1 would almost certainly be too big and expensive to sell well.

I guess the question is, how many potentiels buyers are there, and how much are there willing to put in the PiDP-10?

I would probably put a cap near $500 myself, but what about others? Then if there were an estimated 1000 buyers in the long run willing to spend $500 on it, what replica size would that allow to realistically produce?

Would a crowdfunding and enough interested buyers could help see a 3:4 replica see the light, or is the 2:3 scale the definite sweet spot? I guess that's what a crowdfund could answer.

Damien

--
You received this message because you are subscribed to a topic in the Google Groups "PiDP-10" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/pidp-10/naA2XVfsWvw/unsubscribe.
To unsubscribe from this group and all its topics, send an email to pidp-10+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-10/6b181aa1-3178-6a92-95f7-85d6bcde9592%40quackers.net.

[Richard C]

I am not sure a KL10 would be very interesting. The PDP11 front panel only talked to the Front end PDP11. You could not interact with the PDP10 except through PDP11 RSX11 commands. Also neither KLH10 nor my KL10 simulator fully emulate the Front End PDP11. Mine is only functional, and KLH10 is only enough to talk to CTY.

The 2:3rd replica is easy enough to work with, and use the front panel to interact with the KA10. A KI10 might also be interesting. For software the KA10 gets you Tops 10 4.5-6.03, ITS, Waits, possibly Tenex. KI 10 would get you Tops 6.03-Tops 7.02, possibly Tenex. KL 10 gets you Tops 10 6.03-7.04+, Tops 20 V2-V7, ITS. KS 10 gets you Tops 10 7.01-7,04+, Tops 20 V3 and V4.

I think 2:3rd is the sweet spot. Big enough to be hand soldered yet not so big as to take huge amounts of space which is not really needed.

Rich

[Andrew O'Neill]

Agreed, Richard. I'd be (am!) in the pot for both projects, but surely the starting point must be the one that has the most time and effort devoted to it already! A

You received this message because you are subscribed to the Google Groups "PiDP-10" group.

To unsubscribe from this group and stop receiving emails from it, send an email to pidp-10+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-10/41d165a1-eb34-4b44-8eef-5a39d1d2085en%40googlegroups.com.

[Ethan Dicks]

On Fri, Mar 3, 2023 at 3:05 PM Bill Ezell <w...@quackers.net> wrote:
> So, if a 1:1 was available, would I buy it? Yes, but not sure where I'd put it.

I'd stick mine halfway up the front of an H-960 rack below a
high-speed papertape punch/reader.

-ethan

[pbi...@gmail.com]

I believe that the preliminary working estimate at the current scale was $1000.  But that was before the ongoing global economic bruhaha struck.  I expect that the final cost for this kit is going to be more …

 

From: pid...@googlegroups.com <pid...@googlegroups.com> On Behalf Of Damien Boureille
Sent: Friday, March 3, 2023 3:45 PM
To: PiDP-10 <pid...@googlegroups.com>
Subject: Re: [pidp-10] Crowdfunding the PiDP-10 (and its size issues)

…

I would probably put a cap near $500 myself, but what about others? Then if there were an estimated 1000 buyers in the long run willing to spend $500 on it, what replica size would that allow to realistically produce?

…

Damien

 

[Jim Bailey]

I would definitely be wiling to join a crowdfunding project for this should it be launched :-).  I wouldn't mind ether size, keep in mind his previous kits were also scalled down to something "reasonable".  Most of us don't have a 19" rack in our rooms... yet :-)

[Quinn Cook]

Correct me if I'm wrong... but wouldn't a 1:1 KA10 panel be wider than 19"?

[Eric Smith]

Yes, the KA10 CPU bay with the console is a 31-inch rack.

If anyone needs specific dimensions of the console, I can measure it when I get home.

[tszo...@pacbell.net]

This is a pix of a PDP-10 console.  Note that the switches are buttons not toggles.  It is 29 1/2" wide

[Lars Brinkhoff]

KI10 console in the front, KA10 console in the back.  Very nice!

[Gregory Finn]

The KL-10 CPU power supplies were attached to a swing out door frame.  It was so heavy that it sagged when opened.  Latching it closed was difficult.  An ECO added a small teflon-like plastic ramp at the bottom door casing to help push the sagging door upward.

[oscarv]

Sorry for a slightly slow reply! In my defense, I was spinning the new PiDP-10 PCB and checking the CNC-cut prototype of the coming injection mold, I'm keeping myself busy :-)

Yes, I guess a part of this group would have preferred a full-sized replica. But the cost would be prohibitive, the mold would cost as much as a house. Well, depends where you live...

Even at scale 2:3, it is in fact horrendous. And mind, the cost of a mold goes up with a power of 3 (length, depth, height).

I'm hell-bent to cap the cost of the PiDP-10 at $350, and even that price will be off-putting for those I'd like to reach the most: people new to these machines, but with a curiosity about them. 

Still not sure why I care so much about growing the number of people with hands-on PDP experience, but that seems to be my internal motivator. By the way, even in this group, I feel quite sure that most have no idea of the fascinating stuff that the ITS operating system has waiting for you! I wonder how many people played multi-player Mazewar over the internet and on multiple local terminals. To mention one thing...

Someone in the thread made a very good point about the risk of an injection-molded part this size has of warping, thermal deforming or generally flexing. That's exactly right - it was why I got professional help for this project. It is fascinating to see what tricks they have to avoid these problems. I did not want the machine to feel cheapish, and the PiDP-11 case got away with what it was because you do not actually rest your palms on the switch panel. For the PiDP-10, that is different. You have your hands weighing down on the switch panel. The inner structure of the case is quite complicated, exactly to give it a solid feel. Interesting, learning these injection molding tricks they use. But it won't be my next career choice for sure... fussy.

Anyway - I am glad this project is coming to fruition after three years! 

Kind regards,

Oscar.