Intellivision Keyboard Component pics

[decle]

RonTheCat brought his KC tape deck over today for a bit of cleaning with isopropyl alcohol (don't get it on the belts) and testing. We thought the brotherhood might like a bit of a before:

 

 

and after video:

 

 

Overall electromechanically it seems to be behaving well. The transport has two belts. One is ~19cm x 3mm and connects the main motor to the flywheel, the other is ~8.4cm x 3mm and connects the FF/RW motor to the tape pickup spindles. The pinout of the main power connector when looking at it, pin 1 on the left:

• 1 - Main motor +ve
• 2 - Main motor -ve
• 3 - FF/RW brake? solenoid +ve
• 4 - FF/RW brake? solenoid -ve
• 5 - Playback solenoid +ve
• 6 - Playback solenoid -ve
• 7 - Tape pickup motor +ve
• 8 - Tape pickup motor -ve
• 9 - Not traced

The exact distinction between the purpose of the two solendoids are not really clear at the moment. The one connected to pins 5 & 6 will bring the pinch roller into contact with the capstan. The one connected to pins 3 & 4 will bring the pinch roller forward, but it does not come into contact with the capstan. Both appear to release what is probably a brake on the take up spindles. So it is possible that one solenoid is used for playback, and the other is used for fast forward/rewind. It might also be possible that both solenoids are used in parallel to reduce the load and heating on them. This should become clear when we test the signals being sent to the tape deck by the I/O board.

 

The gunk around the pick up spindle that was turning slowly seems to have been associated with the "brake", we wonder whether it used to have a rubber pad of some description that degraded over time.

 

The fact that all three solendoids work when driven directly, but fail to operate when driven by the KC is a little perplexing, but suggests something might be amiss with I/O board.

 

Overall we have found the tape deck to be less scary to work on than we had thought. If you are competent with a voltmeter, power supply and screwdriver; and are diligent in taking things apart and putting them back togerther, you should be fine.

 

Obviously if anyone has any more information on the deck, how it works and tips they would be greatfully received.

Edited April 28, 2018 by decle

[+nurmix]

RonTheCat brought his KC tape deck over today for a bit of cleaning...

Excellent work and documentation of the cassette unit! I've yet to take one apart and examine it myself, but based on other cassette mechanisms I've seen and worked on, there are an amazing variety of techniques used to drive the various reels and moving bits. From belts and pulleys to tire-drive wheels and brakes, to sprockets and gears. I think you're right about the rubber degrading on the reels, and without those I doubt the deck will be fully functional. Hopefully when I look at the two units I have on loan I can get a better idea of what's there.

 

As for the solenoids getting hot, perhaps the tape head transport needs lubricating, and that's causing them to work too hard moving and keeping it in place?

 

 

 

Sent from my Keyboard Component using Jack's Conversational Intelli-talk cassette

Edited April 29, 2018 by nurmix

[First Spear]

And next... Intellivoice? ECS? System Changer?

[decle]

Hey all,

 

Another bit of an update on Ron's KC. Seems to be coming along nicely, tape drive is responding correctly to keyboard commands. The problem was a blown fuse, doh! We shot the following before replacing the tape door in order to give a clearer view of the mechanism. Then some oaf sticks their fingers in middle of the frame. Oh well!

 

 

We did also try using some electrical shrinkwrap sleeving to replace the degraded "brake pads", but that failed as it caused too much drag on the pick up spools. More thought required.

 

Other bits and pieces we learned:

• The KC is quite happy to operate without either the tape interface board or tape mechanism connected. In both cases it works normally and just reports the cassette cannot be seen if tape operations are requested.
• We are not sure what caused the fuse to blow. Our current best guess is the degraded brake pads that gummed up the mechanism caused the tape head solenoids to draw too much current leading to the blow.
• AV composite mods seem to work with the KC. We did see a strange change in the background colour in the region where the overlay is being applied. This might be down to the rather Frankenstein MC used. This is an NTSC Sears unit that I think has had its EXEC and possibly GROM switched for PAL Inty 1 versions. The background colour is black on startup, but switches to green (as seen) after running a cartridge. This behaviour is also seen on RF. However, an unmodded Inty 1 does not show this behaviour. More investigation required.
• The main drive motors are definitely 12V and driven as such
• The tape eject solenoid is driven with a PWM signal of between 16V and 20V with a duty cycle of ~50% for 0.25sec in order to open the door
• The other two solenoids seem to be driven with a current source. Both show a steady 3V dropped across them when operating and don't get hot. So we were driving them incorrectly last time.
• In addition to the microswitch at the back of the bay the KC also uses the opto-reflector next to the read head to identify the tape state. This is designed to reflect infra-red light of the tape itself. The leader of KC tapes must be transparent to infra-red light allowing the detector to spot the end of the tape.
• The tape sense microswitch can be fooled by shorting pins 1&3 on the small white three pin connector to the small electronics board.
• The timing of David Ahl's BASIC benchmark with the exponent operator (a = a # 2) rather than a = a * a is confirmed to be 98sec.

Next problem is some media...

Edited June 2, 2018 by decle

[JohnPCAE]

Is it possible that the IR is indeed passing through the leader but reflecting off the plastic case behind?

[Intymike]

Wow, this is awesome to watch!

[Ron The Cat]

Felt like Sam Neil seeing Dinosaurs for first time in Jurassic park seeing the tape drive finally ejecting and rewinding from commands typed in on the keyboard

 

Decle is officially my KC service technician going forward

 

Looking in the KC owner's book for the KC it has the name 'Randy' filled in under the Mattel service phone number

 

Seems to indicate the owner must of had to call this person at Mattel when something failed on the KC ,the fuse most probably had blown on the deck power supply board which was soldered onto the board and not something the average consumer could of replaced. The place on the circuit board where the fuse was has a diagram showing a fuse holder was designed to go there like other fuse has on that board but the two connector holes were drilled just a bit too far apart to do that.

Edited June 3, 2018 by Ron The Cat

[intvnut]

IT LIVES! MWA-HA-HA!

 

basic1.jpg

 

basic3.jpg

 

basic2.jpg

 

For giggles we ran David Ahl's BASIC benchmark

 

basic4.jpg

 

Run time was 62 seconds. Pretty respectable for a 0.89MHz 6502, that puts it about half way down the list of machines (see below), in the company of some reasonable business machines and well ahead of its contemporaries like the C64 and Apple II. This performance does seem to come at the cost of some floating point accuracy, although it is about the same as some other machines with MS Basic like the TRS-80 Model II and Aquarius. It should be noted that if you try to replicate this in MAME you will get a time of about 30 seconds, I think it runs the 6502 at twice its real clock speed.

 

 

 

 

 

 

I'm rather curious why the KC BASIC comes out so much faster than the Apple II. It doesn't quite make sense given what I know about its BASIC.

 

I suspect it's the fact you sub'd in RND(0) instead of RND(1), and changed A^2 to A*A. RND(0) returns the previous random number, while RND(1) generates a new random number. Also, "A*A" does something different than "A^2". The latter is actually EXP(LOG(A) * 2), and so should be rather more expensive. According to this, you can use # in place of the missing ^ symbol. That perhaps explains why your precision doesn't match any of the other BASICs in the chart.

 

The only real major difference between Applesoft II and Keyboard BASIC is 6-digit vs. 9-digit float, according to Michael Steil's analysis. That, and the CHRGET routine is much slower on the KBD: https://www.pagetable.com/?p=46

 

I'd expect the KC to be about the same speed as Applesoft II: Slightly slower due to the clock rate and CHRGET; slightly faster due to the reduced precision. I'm not sure how much faster due to 6-digit, as the only other 6-digit 6502 machine is the OSI Challenger, but it uses a version of BASIC that replaces ROR with a really expensive sequence (a workaround for an early, buggy version of the 6502), so that explains why its benchmark is waaay off.

[+DZ-Jay]

Hi, Decle,

 

I do not know if the tapes expected were actually "special Keyboard Component tapes." As I recall back in those days, there were cassette tapes that were sold specifically for "data" rather than music. I never understood what would be the difference back then, but I do remember having cassette tapes with absolutely clear and transparent lead. Perhaps that is one of the differences, since infra-red sensors to detect tape lead seems to have been a common thing back then.

 

Perhaps it's a matter of finding the right kind of tape, which may not be too hard.

 

-dZ.

[+DZ-Jay]

 

 

I'm rather curious why the KC BASIC comes out so much faster than the Apple II. It doesn't quite make sense given what I know about its BASIC.

 

I suspect it's the fact you sub'd in RND(0) instead of RND(1), and changed A^2 to A*A. RND(0) returns the previous random number, while RND(1) generates a new random number. Also, "A*A" does something different than "A^2". The latter is actually EXP(LOG(A) * 2), and so should be rather more expensive. According to this, you can use # in place of the missing ^ symbol. That perhaps explains why your precision doesn't match any of the other BASICs in the chart.

 

The only real major difference between Applesoft II and Keyboard BASIC is 6-digit vs. 9-digit float, according to Michael Steil's analysis. That, and the CHRGET routine is much slower on the KBD: https://www.pagetable.com/?p=46

 

I'd expect the KC to be about the same speed as Applesoft II: Slightly slower due to the clock rate and CHRGET; slightly faster due to the reduced precision. I'm not sure how much faster due to 6-digit, as the only other 6-digit 6502 machine is the OSI Challenger, but it uses a version of BASIC that replaces ROR with a really expensive sequence (a workaround for an early, buggy version of the 6502), so that explains why its benchmark is waaay off.

 

Whoa! Whoa! I just got a huuuuuuge dose of deja vû. I kinda seem to remember you posting something about the precision of the A^2 before, perhaps in a different thread? If you didn't, then I am probably just going crazy... or remembering the future?

 

In any case, fascinating stuff!

 

-dZ.

 

 

EDIT: Corrected to say "remember you posting..." I stupidly forgot the "you," making it sound like I'm taking credit, sorry.

Edited June 3, 2018 by DZ-Jay

[intvnut]

• AV composite mods seem to work with the KC. We did see a strange change in the background colour in the region where the overlay is being applied. This might be down to the rather Frankenstein MC used. This is an NTSC Sears unit that I think has had its EXEC and possibly GROM switched for PAL Inty 1 versions. The background colour is black on startup, but switches to green (as seen) after running a cartridge. This behaviour is also seen on RF. However, an unmodded Inty 1 does not show this behaviour. More investigation required.

 

The EXEC version shouldn't matter, as the code that differs between Sears and Mattel hasn't run yet. The GROMs are identical between Sears and Mattel, FWIW.

 

The KC EXEC sets the display to green by programming the border color register only:

.

;;=============================================================================
;;  ISR:  Make display green.  (Initial powerup ISR.)
;;-----------------------------------------------------------------------------
L7726:  MVII    #$0005, R0      ; $7726  \
L7728:  MVII    #$002C, R3      ; $7728   |__ Make display green.
        MVO@    R0,     R3      ; $772A   |
        MOVR    R5,     R7      ; $772B  /

.

 

This implies that the KC EXEC expects the STIC to be blanked. How weird...

[intvnut]

 

Whoa! Whoa! I just got a huuuuuuge dose of deja vû. I kinda seem to remember you posting something about the precision of the A^2 before, perhaps in a different thread? If you didn't, then I am probably just going crazy... or remembering the future?

 

In any case, fascinating stuff!

 

-dZ.

 

 

EDIT: Corrected to say "remember you posting..." I stupidly forgot the "you," making it sound like I'm taking credit, sorry.

 

 

It wasn't me, it was mr_me. I was catching up on this thread, and didn't see his post yet.

[+DZ-Jay]

 

 

It wasn't me, it was mr_me. I was catching up on this thread, and didn't see his post yet.

 

Ah! OK, I feel better. It really felt a bit weird, as if I couldn't be sure if it was a real memory or just fake "deja vû." On the downside, I guess I do not have powers to remember the future, then.

[1980gamer]

 

Ah! OK, I feel better. It really felt a bit weird, as if I couldn't be sure if it was a real memory or just fake "deja vû." On the downside, I guess I do not have powers to remember the future, then.

Didn't you say that before? Hmmm, could swear I remember this...

 

Monday morning, when I get to work... I suffer from vû ja-de.

That's the feeling you will do this again.

[+Lathe26]

Hey all,

 

Another bit of an update on Ron's KC. Seems to be coming along nicely, tape drive is responding correctly to keyboard commands. The problem was a blown fuse, doh! We shot the following before replacing the tape door in order to give a clearer view of the mechanism. Then some oaf sticks their fingers in middle of the frame. Oh well!

 

https://www.youtube.com/watch?v=wNt95D-A5Ks

 

 

Awesome work! Very informative and well explained what has been uncovered.

 

The lovely framed photo in the background is fantastic! ??

[decle]

Preparations for the next visit to RonTheCat's lab are coming along nicely...

 

 

...the question is, will it all work?

 

Is it possible that the IR is indeed passing through the leader but reflecting off the plastic case behind?

 

This is something we tested on the day. If we put the tape shell in the drive, but pulled the tape itself out and passed it behind the sensor, the drive correctly interpreted this as the beginning of the tape. So it looks as though the leader on the standard audio tape, whilst translucent to visible light, might reflect some infra-red. We also know that infra-red light is used as we used the well known sensitivity of phone camera CCDs to IR to see when the LED was operating (all the time the tape is active). As a side note VCR's used a similar light / IR mechanism to dectect the end of tapes.

Edited July 21, 2018 by decle

[+DZ-Jay]

Preparations for the next visit to RonTheCat's lab are coming along nicely...

 

prep1.jpg

 

...the question is, will it all work?

 

 

This is something we tested on the day. If we put the tape shell in the drive, but pulled the tape itself out and passed it behind the sensor, the drive correctly interpreted this as the beginning of the tape. So it looks as though the leader on the standard audio tape, whilst translucent to visible light, might reflect some infra-red. We also know that infra-red light is used as we used the well known sensitivity of phone camera CCDs to IR to see when the LED was operating (all the time the tape is active). As a side note VCR's used a similar light / IR mechanism to dectect the end of tapes.

 

Er... I'm curious to know what sort of support function do those Lego panels serve...

 

-dZ.

[mr_me]

You should make a KC tape from the backups that Frank P has.

[+JasonlikesINTV]

Preparations for the next visit to RonTheCat's lab are coming along nicely...

prep1.jpg

 

 

Excellent example of knolling Edited July 21, 2018 by JasonlikesINTV

[+Lathe26]

Preparations for the next visit to RonTheCat's lab are coming along nicely...

 

prep1.jpg

 

...the question is, will it all work?

 

 

This is something we tested on the day. If we put the tape shell in the drive, but pulled the tape itself out and passed it behind the sensor, the drive correctly interpreted this as the beginning of the tape. So it looks as though the leader on the standard audio tape, whilst translucent to visible light, might reflect some infra-red. We also know that infra-red light is used as we used the well known sensitivity of phone camera CCDs to IR to see when the LED was operating (all the time the tape is active). As a side note VCR's used a similar light / IR mechanism to dectect the end of tapes.

 

Cool! I know what the 2 purple boards with the black connectors are, but I am curious as to the R-Card shell and the cartridge board with the tiny daughterboard.

[First Spear]

IR sensor and microswitch to detect a special tape with super-clear header area?! Just another example of places for the unit to be unreliable.

 

You guys are doing great, great, great work. Thank you for sharing! It is now easy to see why these never really made it, they were more of an engineering-driven project than a marketing-driven project; I can imagine that if a business unit leader said "let's make this" instead of Papa Intellivision, it would have had 1/2 the features, been more reliable, and sold. Wow, the KC was just too complex to survive as a real product.

 

 

Hey all,

 

Another bit of an update on Ron's KC. Se[snip]

[+DZ-Jay]

IR sensor and microswitch to detect a special tape with super-clear header area?! Just another example of places for the unit to be unreliable.

 

You guys are doing great, great, great work. Thank you for sharing! It is now easy to see why these never really made it, they were more of an engineering-driven project than a marketing-driven project; I can imagine that if a business unit leader said "let's make this" instead of Papa Intellivision, it would have had 1/2 the features, been more reliable, and sold. Wow, the KC was just too complex to survive as a real product.

 

 

 

Hmm... I don't quite think so. The reason it was unreliable was because of trying to keep costs down. Most of the features implemented already existed in one way or another on professional/enterprise/high-end level hardware; the idea was to provide something similar at a fraction of the cost using consumer-grade parts.

 

Remember, those features were driven by marketing, not engineering. It was intended to be a "full-fledge professional computer" that stood out from the typical micro-computer-with-a-datassette "toy."

 

The tape-lead detection mechanism wasn't some cobbled-together hack they came up with; it was standard in many devices of the time, and later. If anything, the Keyboard Component was just ahead of its time as a consumer product: there was no way to build something like that and keep it under several $1000.

 

By the time something like it could have been done practically, data storage was moving towards disk-based rather than tape-based, which already provided random-access seek built-in.

 

You are right that the reason the project continued for so long was due to Papa Intellivision's influence and clout within the company; but I do not believe anything less than the actual Keyboard Component would have made any impact to in the market (already getting saturated with cheap micro-computers that doubled as game consoles).

 

Oh, and a "business unit leader" did say "let's make this" instead: It was the ECS, which in my personal estimation, is an embarrassment to be called a "computer" -- and this is coming from someone who owned one when he was a young impressionable child, as his very first computer (my father sold it to a neighbor within the first year and we got a C=64).

 

-dZ.

[decle]

...the question is, will it all work?

 

No, no it won't

 

Er... I'm curious to know what sort of support function do those Lego panels serve...

 

I'm afraid the purpose of the Lego will remain a secret for a while longer...

 

Cool! I know what the 2 purple boards with the black connectors are, but I am curious as to the R-Card shell and the cartridge board with the tiny daughterboard.

 

This post might help with the identity of the mystery board.

 

In the mean time Ron and I have done a little more investigation of the tape drive. This is the circuit on the bottom of the drive unit:

 

 

 

From these pictures we have put together the following partial schematic:

 

audioBoard.txt <--- rename to .sch to load into Eagle

 

Couple of things to note. The values in the diagram, including the transistor types, are taken from the images, and so are not always visible. As such, they should be taken with a large pinch of salt. At this point I should also confess that my analogue electronics knowledge is rusty as hell. In looking at this circuit I have managed to convince myself that either side could handle the audio or digital tracks. This write up is primarily to drive testing at the next visit to Ron's lab when we should be able to firm up the details.

 

That said, my current thinking is:

 

I suspect the majority of the left half of the board is probably related to the handling of the digital tracks.

 

It looks as though E2 might be the read only digital track, E3 the read / write digital track. With the 4066B controlling whether the signal out on JP1-1 is routed to the head or which of E2 and E3 are routed to the amp and from there to JP1-5 and JP1-6. It looks as though both E2 and E3 can be read concurrently. It is also possible to route both E2 and E3 to IC2A, but only E3 can be routed to IC2B. The reason for this is not clear. The two amplifiers primarily look to be set up as voltage followers with a gain of 4.3. I'm not sure what the inclusion of the capacitors and Q999 is doing, perhaps a filter or output control (hey I said my analogue was rusty )?

 

In addition, there is a small circuit to handle the tape leader sensor. The IR-LED is constantly fed from +ve via E7 and the photo-diode is then fed back through E8 to JP1-8.

 

So the full set of signals on JP1 is:

• JP1-1 - digital data write
• JP1-2 - digital write enable (connects JP1-1 to E3)
• JP1-3 - digital read enable (connects E2 to IC2A and E3 to IC2B)
• JP1-4 - weird digital read enable (connects E3 to IC2A)
• JP1-5 - digital read only data (IC2A output)
• JP1-6 - digital read / write data (IC2B output)
• JP1-7 - unknown - IC2A output control?
• JP1-8 - tape header light sensor
• JP1-9 - +ve supply

This would suggest that the right half of the board handles the audio. This based on the output from IC3A and IC3B being merged and fed out on a single line JP2-5 and the copious amounts of AC coupling. I think it only makes sense to merge the audio tracks in this way. If correct this would mean that:

• E9 / E10 are connected to the read / write audio track
• JP2-1 controls the read / write of the audio track
• E11 / E12 are connected to the read only audio track
• It looks as though JP2-3 and JP2-7 might suppress the output of each of the audio tracks to JP2-5. However, my op-amp analysis is too rusty to be sure. If this is not what Q4 and Q6 do I don't think it is possible to isolate the playback of the read only track (i.e. if user audio was recorded at the same point in time as a piece of pre-recorded audio it would not be possible to playback the user audio without also hearing the Mattel audio).
• Again, I'm guessing the CR chains in the op-amp feedback loops create filters.

In addition, on this half of the board there is control for the erase head (E13 / E14, active low) and the eject solenoid passthrough (E15 / E16)

 

So in terms of JP2 signals we have:

• JP2-1 - !Read / Write - when low holds E9 low and allows E10 to connect to IC3A
• JP2-2 - IC3A read enable?
• JP2-3 - Audio signal record
• JP2-4 - Ground
• JP2-5 - Audio signal play
• JP2-6 - IC3B read enable?
• JP2-7 - Erase head enable (active low)
• JP2-8 - Eject solenoid signal
• JP2-9 - Eject solenoid ground

As I said my analogue electronics is very rusty from 25 years of neglect, therefore, any corrections or insights would be gratefully received.

 

Edited August 6, 2018 by decle

[Jeffrey Bouchard]

This is cool! Does this mean that emulation is becoming easier or a possibility soon?

[+DZ-Jay]

This is cool! Does this mean that emulation is becoming easier or a possibility soon?

 

I thought there was already emulation available. It just is useless without the data from the tapes or carts... I could be wrong, though, but I believe Joe Z. once said he had KC emulation ready.