PDA

View Full Version : CT-1024 Terminal Manual?



falter
February 23rd, 2014, 10:15 PM
I think this one will be directed at Chuck mostly.. Chuck do you know if there was an operation manual for the ct-1024? I've looked all over Matt Holleys site.. he has assembly instructions that give a bit of info on starting the unit up for testing.. but not much else. I'm trying to figure out how to run it in just tv typewriter modr.. to see if the keyboard works. Im getting replacement capacitors.. and i dont think the serial board is working.. but there is a switch that brings up a page of ? marks and I seem to recall reading *somewhere* that there was a keystroke or button or something that prepared the unit for typing. I just want to see the extent of functionality if I can.

One other question... off one of the leads from the power supply I'm getting 12.95v.. is that likely tied into the old capacitor issue? I'm hoping i've ordered the right replacements. All the originals were measured in mfd.. had tobuse a calculator to try and figure out the uf equivalents.

Doug G
February 24th, 2014, 01:58 PM
As far as I know 68mf = 68mfd = 68uf

When I started in ham radio in the late 50's caps were commonly shown and labeled as mfd rather than the newer uf symbol.

Chuck(G)
February 24th, 2014, 02:54 PM
Nope, as far as I remember, you turned it on and hooked it up. The most helpful part of the manual was the schematic for me.

Jumper (RxD and TxD and you should be in local mode.

falter
February 24th, 2014, 06:26 PM
I tried that but didn't get anywhere. I'm guessing most of my problems are in the power supply. I temporarily removed the serial board and plugged the keyboard into the port on the motherboard itself, but that achieved nothing either, except when I was playing with the voltages I briefly got it to come up to a field of Ns, and if I hit a key it changed them to whichever key I pressed.

Hoping I get those caps soon and that that helps my problem out a bit. I'm having trouble finding the big 5000uf 10v one though.. all I can find thus far are 'axial'.

Chuck(G)
February 24th, 2014, 06:30 PM
The standard PSU for the CT1024 was pretty awful. I rebuilt it from scratch--the TVT draws a lot of power for which the original PSU was woefully under-designed. Today, you can get a pre-made power supply to do the same thing for not much money.

falter
February 24th, 2014, 10:35 PM
I kind of thought about getting an external psu.. so as not to take away from the original unit (I'm thinking I could just connect it temporarily when I want to use it). But I'm intimidated by that kind of thing.. no idea what to buy.

Chuck(G)
February 25th, 2014, 09:28 AM
Okay, let me have a look at my documentation and I'll get back to you if no one else does.

falter
February 26th, 2014, 08:11 AM
This is what I found for 5000uf 10V. But it says 'screw type'. Only other one I found was axial. Seems to be a tough one to locate.. but that's the biggest cap on that power supply.

http://ca.mouser.com/ProductDetail/Cornell-Dubilier/3186BD502U010APA1/?qs=sGAEpiMZZMukHu%252bjC5l7YVVJddEcKE%2fzeGuGaDx5 YiM%3d

Chuck(G)
February 26th, 2014, 08:32 AM
There's no particular reason that it has to be screw terminal. In the days of "brute force" linear DC supplies "Computerlytic" caps were very common. Nowadays with switching supplies, they're somewhat uncommon. I suspect a "snap-in" radial cap would do just fine.

At any rate, I found my manual for the P197 PSU used for the CT-1024. The ratings stated are +5V, 2.25A, -5V, 20 ma, and -12V, 60ma. So a pretty limp-wristed supply. FWIW, the RS-232 outputs are also pretty limp-wristed at +/- 4.75V. A 3-ourput PSU with those ratings should be pretty simple to find.

falter
September 26th, 2014, 04:05 PM
Hey Chuck,

I'm just getting onto this now.. is this sort of the thing you were talking about: http://www.ebay.ca/itm/B780-ELECTRO-INDUSTRIES-TRIPLE-OUTPUT-POWER-SUPPLY-2555A-VARIABLE-DC-/390929146062?pt=LH_DefaultDomain_0&hash=item5b053160ce

I'm just not sure about the -5V and how that all works. I've never used one of these power supplies. Probably have the wrong one. :)

Chuck(G)
September 26th, 2014, 04:44 PM
Oh, good heavens no! Nothing that fancy (although a good bench supply can be useful :) )

Even a small (mini ITX) PC power supply should do. Or use one of the surplus ones that seem to be wandering around. Did you ever find a good surplus store in BC? Otherwise, there's Jameco.

falter
September 26th, 2014, 05:44 PM
Oh, good heavens no! Nothing that fancy (although a good bench supply can be useful :) )

Even a small (mini ITX) PC power supply should do. Or use one of the surplus ones that seem to be wandering around. Did you ever find a good surplus store in BC? Otherwise, there's Jameco.

Yes I was considering Jameco. I think the problem is the supply your own power supply thing predates me and I feel really nervous about getting the wrong thing. I'm concerned about making sure the amperages work, etc. I understand they can be more than you require but not less?

Chuck(G)
September 26th, 2014, 06:07 PM
The current, yes--it can be much lower. That's why I posted the CT1024 specified ratings.

daver2
September 27th, 2014, 12:21 AM
Falter,

Yep, that power supply should do - with the proviso that the three outputs are isolated from each other and ground. I found some additional documentation at http://www.electroindustries.net/specifications.php#2555A which seems to state that is the case.

I would connect as follows:

Set the switch to INDEPENDENT (so that you can adjust the A and B voltages separately).

Set the 'A' side for 5V.
Set the 'B' side for 12V

Join the + side of the 'A' supply to the + side of the 'B' supply to the - side of the fixed 5V supply. This is your common 0V lead. Also connect to GND post.

The - side of the 'A' supply is now your -5V supply.
The - side of the 'B' supply is now your -12V supply.
The + side of the fixed 5V supply is your +5V supply.

The fixed +5V supply is capable of delivering 3A (which is greater than the 2.25A that Chuck has identified). The 'A" and 'B' supplies are each capable of delivering up to 0.5A (which is greater than the 20mA and 60mA that Chuck has identified).

I would, however, check the voltages with a multimeter before connecting them to you CT-1024 just to be sure you have wired everything correctly. Connect the -Ve lead of your multimeter to the common 0V lead from the power supply (GND) - and measure each voltage from the outputs separately. 'A' should indicate -5V, 'B' should indicate -12V and the fixed supply should indicate +5V.

I am looking for a new bench supply - that one looks good apart from the cost of shipping and import duty to the UK (and the seller states that they only ship to the States and Canada).

Dave

falter
September 27th, 2014, 07:30 AM
Thanks muchly Dave! Your detail is exactly what I need being new to bench power supplies, etc. So when you say 'join', you mean literally there will be wires linking all four terminals ( A +, B +, 5v -, gnd)?

Another question.. do I need to remove the old power supply completely, or could I just clip my leads to the voltage outputs on its pcb?


Falter,

Yep, that power supply should do - with the proviso that the three outputs are isolated from each other and ground. I found some additional documentation at http://www.electroindustries.net/specifications.php#2555A which seems to state that is the case.

I would connect as follows:

Set the switch to INDEPENDENT (so that you can adjust the A and B voltages separately).

Set the 'A' side for 5V.
Set the 'B' side for 12V

Join the + side of the 'A' supply to the + side of the 'B' supply to the - side of the fixed 5V supply. This is your common 0V lead. Also connect to GND post.

The - side of the 'A' supply is now your -5V supply.
The - side of the 'B' supply is now your -12V supply.
The + side of the fixed 5V supply is your +5V supply.

The fixed +5V supply is capable of delivering 3A (which is greater than the 2.25A that Chuck has identified). The 'A" and 'B' supplies are each capable of delivering up to 0.5A (which is greater than the 20mA and 60mA that Chuck has identified).

I would, however, check the voltages with a multimeter before connecting them to you CT-1024 just to be sure you have wired everything correctly. Connect the -Ve lead of your multimeter to the common 0V lead from the power supply (GND) - and measure each voltage from the outputs separately. 'A' should indicate -5V, 'B' should indicate -12V and the fixed supply should indicate +5V.

I am looking for a new bench supply - that one looks good apart from the cost of shipping and import duty to the UK (and the seller states that they only ship to the States and Canada).

Dave

MikeS
September 27th, 2014, 07:31 AM
Sheesh! Just use an old PC supply.

Chuck(G)
September 27th, 2014, 09:23 AM
Sheesh! Just use an old PC supply.

What I was trying to say.

falter
September 27th, 2014, 10:02 AM
Okay.. so I have the original psu out a bit:

20676

20677

So if I understand correctly, I can attach the pins for 5v, -5v etc of the AT psu by wire to the appropriate terminals on this old power supply (I am reluctant to desolder it). The only thinga that throw me are where the ground wires connect up (they appear to be a fairly heavy gauge) and that white wire that is soldered in beside one of the resistors.

Chuck(G)
September 27th, 2014, 10:35 AM
Sorry, your illustrations don't show where the other end of the white wire goes. I suppose I could check my assembly instructions, if it's not possible to show that.

EDIT

Forget it--that white wire is the 60Hz reference from the transformer secondary--not shown on the schematic--nominally 6.3VAC.

Foo--it might be simpler just to refurb and beef up your current supply. The big problem is the TO-220 MJE3055 transistor. It needs to pass a higher current than the package and heatsink is rated for. Perhaps using a TO-3 2N3055 on a substantial heatsink might work better. Replacing the electrolytic capacitors probably wouldn't hurt either.

falter
September 27th, 2014, 11:37 AM
Yeah. I had a feeling this wasn't going to be that easy. And I am being *very* cautious.. this is the oldest piece in my collection (all the chips date to 75) so I don't want to make things worse, and I'm also keen to keep it as close to original as possible.

And you're fairly confident we're chasing the right suspect here with PSU right? The original problem was video like this:

https://www.youtube.com/watch?v=2WQRCQiXRcE

And no response from keyboard. I'm still looking at the docs to figure out what the switches do.. there is one that switches between that field of question marks and nothing. There's also that very brief line of garbage characters at the beginning. I've wondered if part of this has to do with the monitors I'm using vs. what they were expecting you to connect this to back in the day..

Chuck(G)
September 27th, 2014, 12:53 PM
Looking at the display, you have serious hum/AC ripple problems.

The switches that you have don't match anything on mine. You should have momentary contact pushbuttons for clear screen, line, etc.

So let's go back to zero on this.

First off, what shape (and in particular, the +5V output) is your power supply in? How much ripple do you detect?

Second, remember that the keyboard and display sections are completely separate. I assume that you have a UART card in there with an AY-4-1013, S1883 or TR1602 IC on it? It's completely normal for the screen to come up in garbage--there's no power-on preset for the display RAM. You should be able to clear it with the pushbutton that goes to pin 8 on J3. Similarly, you should be able to select which display page you're showing with the switch that grounds either J10 pin 1 or pin 2.

Similarly, the pushbutton that goes to J3, pin 2, is "forward cursor"; to pin 4, "backward cursor"; to pin 1, "home"; to pin 5, "line up"; and to pin 3, "line down". You might want to label them.

A little tip--you can significantly reduce the power consumption of the terminal by replacing the 6-2102 ICs on the display memory board with 21L02 or 91L02 chips. Those old 2102s devices and get pretty hot.

falter
September 27th, 2014, 02:18 PM
I don't know how to check for ripple.. but I did do a multimeter test off the +5V pole, and it is steady at exactly 5v. The other two are harder to get at but I'll try.

The pushbuttons appear to do nothing.. except one of them does sort of cause a 'reset' every once in a while.. but the screen returns to garbage.

And yes, my system has the optional serial board installed.

Chuck(G)
September 27th, 2014, 02:41 PM
Did you loop back your serial (keyboard) output to the (display) input? That is, JS-1 pin 6 to pin 7.

Put your DMM between JS-1 pin 6 and ground? Do you see a pretty high negative voltage? If you press a key on the keyboard, do you see that voltage twitch?

To check for ripple, set your DMM to "AC volts", and then see if you get any reading across the various terminals. That big smear in the display is an indicator.

falter
September 27th, 2014, 02:51 PM
I'm using an analog MM right now.. I set the switch to ACV 10 -- and when I attach to the 5v+ pole, it goes to +22db?

falter
September 27th, 2014, 02:58 PM
Also, I may have goofed on the loopback.. I looked at the rs232 port on the back, found pin 6 and 7 and jumpered them.. but looking inside.. it appears the previous owner has this wired differently -- 7 has a wire to it but there's no wire to 6.

Chuck(G)
September 27th, 2014, 03:10 PM
Well, in this case a DMM is probably best. On the DB25, connect pins 2 and 3.

falter
September 27th, 2014, 03:11 PM
Okay, found the wire coming from pin 6 on JS-1 (if I understand correctly, you count from top left, left to right, from top to bottom.

I get -12.83V. No change at all with a key pressed.

falter
September 27th, 2014, 03:15 PM
Note: on the first test, I did not have 2 and 3 on the DB25 jumpered. When I connect pins 2 and 3 on the DB25 and then read voltage from JS-1 6, it comes down to -5.85V. But again, no change with pressing keys.

falter
September 27th, 2014, 03:34 PM
I tried removing the cursor board and shorting J3 pin 8, also J3 pin 2 and pin 1 per the assembly instructions. No reaction at all. I'm kind of wondering about this one switch that when flipped causes the entire field of characters to go away until you flip back again.

Chuck(G)
September 27th, 2014, 03:54 PM
Well, I don't know what the switch connects to, unless it's to J10 1-2.

You can also ground pin 8 of JS-1 on the serial board to echo the keyboard characters (that bypasses the level shifting circuitry).

Do you perhaps own a logic probe or oscilloscope? A VOM is not much of a tool to do trouble-shooting with, particularly on something this old.

falter
September 27th, 2014, 04:10 PM
Okay I think I solved that switch (and am solving the rest).

The mystery switch that causes the characters to disappear connects to pin 6 of 10 (Cursor ON/OFF).

There is a second mystery switch that goes to pin 5 of the Serial board's I/O connector.

A third switch, right below it, connects to pin 8 of the serial board I/O connector.

falter
September 27th, 2014, 04:20 PM
Answered my own question by re-checking JS-1 pinouts.

The top switch goes to pin 5 of JS1 which is Ground for Receiver OFF
The bottom switch goes to pin 8 of JS1 which is Ground for Echo OFF.

So flipping the top switch to where both its poles have contact means Receiver OFF, I take it? Same for Echo?

So if Echo OFF is grounded, nothing will come back to the screen when I type..

Chuck(G)
September 27th, 2014, 04:27 PM
I'm assuming that the keyboard is plugged into JS-2 on the serial interface board. But yes, Pin 8 of JS1 should not be grounded for echo operation.

falter
September 27th, 2014, 04:35 PM
Yes it's plugged into the connector marked 'Keyboard' on the serial board which I assume is JS-2.

Okay, so I think I sorted out my push buttons:

Button 1 - Pin 3 of JMC 1 = Cursor Down
Button 2 - Pin 8 of JMC 1 = Ground
Button 3 - Pin 9 of JMC 1 = Home Up
Button 4 - Pin 5 of JMC 1 = Erase to EOL
Button 5 - Pin 4 of JMC 1 = Erase to EOF

Big green stalk switch = Page 1 or Page 2

The dial must be for baud.. so I think I've got all my switches figured out. Now if only I understood how they worked.. :)

falter
September 27th, 2014, 04:44 PM
20682

When the cursor switch is on.. that is when the garbage screen disappears. But I've tried hitting cursor down a bunch of times and typing.. nothing. I suspect hidden up top is that strange single line of random characters.

falter
September 27th, 2014, 05:31 PM
By the way.. did that +22db for the ac noise test confirm what yoh suspected?

Chuck(G)
September 27th, 2014, 05:49 PM
No, I was after a volt- or millivolt reading--I don't know what your VOM is using as a reference; it could be somewhere around 200 mV pp, but that's hard to say. What reading do you get when you measure 6.3VAC RMS?

falter
September 27th, 2014, 07:23 PM
This is the DMM I have.. not really sure how to test as you suggested. ;) I've gotten better theblast year with continuity, resistance, etc.. but we are into uncharted territory for me with AC hum, etc. ;) If you dont mind guiding me a bit I'm sure I'll catch on (or catch fire :)).

20685

I shoulf mention under the red connector its marked AC, and under the black.. V with the OHMs symbol.

Chuck(G)
September 27th, 2014, 08:02 PM
It looks as if the only AC scales on the DMM are for 200 and 750 V. Fine for the home electrician, but not so much for digital circuits.

An inexpensive logic probe may help you. It can detect the difference between a steady level and a pulse.

But honestly, I'd probably not jump into this thing without an oscilloscope. Then it'd probably be a matter of a couple of hours of checking.

falter
September 28th, 2014, 01:03 PM
It looks as if the only AC scales on the DMM are for 200 and 750 V. Fine for the home electrician, but not so much for digital circuits.

An inexpensive logic probe may help you. It can detect the difference between a steady level and a pulse.

But honestly, I'd probably not jump into this thing without an oscilloscope. Then it'd probably be a matter of a couple of hours of checking.

Yeah I kind of figured. It'll be a while yet before I have a scope and so much of this is yards above my head. It'll take a long time to get there.

Is there any point to just trying cap replacement? I'd be up for that.. although I cannot for the life of me find a 5000mfd 10v can type with just two legs.

Chuck(G)
September 28th, 2014, 02:38 PM
Is there any point to just trying cap replacement? I'd be up for that.. although I cannot for the life of me find a 5000mfd 10v can type with just two legs.

It wouldn't hurt. Electrolytics have been getting smaller and smaller (better materials and processes.) So, an axial-lead 10,000 uF, 16V capacitor (http://www.jameco.com/webapp/wcs/stores/servlet/Product_10001_10001_609473_-1) (which exceeds the original by almost a factor of 2--it won't hurt) is, in a radial-lead package, only 18 mm in diameter by 41 mm tall. (And they can be had in smaller packages).

If you're trying to reproduce the "look" of old equipment, that can be a problem. I've seen people "gut" the old can and solder a new capacitor inside and then reassemble the old one. Called "capacitor stuffing" (google for it)--it's something that goes on a lot in the antique radio field.

falter
September 28th, 2014, 03:33 PM
It wouldn't hurt. Electrolytics have been getting smaller and smaller (better materials and processes.) So, an axial-lead 10,000 uF, 16V capacitor (http://www.jameco.com/webapp/wcs/stores/servlet/Product_10001_10001_609473_-1) (which exceeds the original by almost a factor of 2--it won't hurt) is, in a radial-lead package, only 18 mm in diameter by 41 mm tall. (And they can be had in smaller packages).

If you're trying to reproduce the "look" of old equipment, that can be a problem. I've seen people "gut" the old can and solder a new capacitor inside and then reassemble the old one. Called "capacitor stuffing" (google for it)--it's something that goes on a lot in the antique radio field.

So you can use a capacitor that is rated higher than the original? I kind of wondered why they were so small. Capacitor stuffing sounds interesting.

I'm torn because this piece as it is is pretty much 100% original. I don't have anything that matches it for age or uniqueness. I also wonder if I'd ever really use it. If I got a SWTPC 6800 I might want to uave the option. But I might just prefer a Lear-Siegler.

My gut feeling is we have some bad ics in there. It might be semi usable for demo purposes if I could nail down which. What to do.. what to do.

Chuck(G)
September 28th, 2014, 05:16 PM
Well, the CT1024 was a little on the crude side. Right away, I knew that I'd need software control of the special funcitons, such as cursor up, down, erase, etc. I wired up a small prototype board, interestingly with passive (i.e. diode) logic to provide that. I also turned it into a single-page 80 column display (not too difficult). Even so, it was uppercase only and pretty crude compared to "real" terminals.

However, about a year later, I went to yet another "single board" terminal that was much superior, cheaper, and one that had most of the features I needed.

Eventually, I went to a Beehive SuperBee, which was a real joy by comparison.

daver2
September 29th, 2014, 10:57 AM
I am now getting a little confused about where you are.

Are you using the original power supply or a PC one at the moment? If you are using the original power supply then (on the assumption that the capacitors are drying up) you could just temporarily add some new capacitors in parallel with the originals as a test. Don't forget to do it on all three power rails just to be safe. You definitely need a rock-solid power supply otherwise you may be chasing your tail with the logic! If you are using a PC supply - what have you done with the 6.3V a.c. reference signal? This signal is used to derive the synch pulses and without it you will almost certainly not get a locked video display.

J10 pin 6 (marked on the schematic I have as J10 pin 9 for some reason) is a CURSOR BLINK ON/OFF not a CURSOR ON/OFF. This switch should select whether the cursor is blinking or steady. IC8 is an NE555 acting as an astable and controls the speed of the cursor blink. Grounding J10 pin 6 should cause the oscillator to stop oscillating. If you place an LED with a series current limiting resistor between +5V and the collector of Q2 (or use an oscilloscope/logic probe) you should see the cursor blink when J10 pin 6 is open circuit and a permanent 0V/GND with J10 pin 6 grounded.

The CURSOR can be disabled totally by wiring pin 'M' to ground.

Changing the switch on J10 pin 6 should not cause characters to disappear (assuming the power supply is 'good').

IC1 is another NE555 timer and is used in conjunction with the 6.3V a.c. mains reference voltage to derive the horizontal frequency. There is an excellent write-up in http://www.swtpc.com/mholley/CT_1024/CT1024_Assembly.pdf about half-way down. My gut feeling is that you will probably need an oscilloscope to do much in the way of troubleshooting unless it is something as simple as a permanently duff gate/IC (i.e. the signal is either present or not present).

Dave

falter
September 29th, 2014, 07:49 PM
Thanks Dave.

There is no cursor at all no matter what I do. The only thing that changes at all is when switch the cursor button and the screen goes blank (well, not blank, there's that messed up top bar but you can't read it).

Sounds to me like power is definitely a good suspect. You can see in the video -- sometimes it's like it is there.. you turn it on and there's a long pause before anything appears, and then sometimes it's there instantly. And there's that weird random line with messed up characters.

I'm just a bit hesitant about taking it apart. This thing is special to me because of it's age and uniqueness, even though it's just a terminal. Because I know nothing about power supplies (but am working on that), I wouldn't feel confident getting into it unless someone was able to say to me -- yes, replace that cap with that, replace that cap with this modern equivalent) etc etc. This is a problem dogging me with my replica project too.. not knowing what certain transformers are, etc.

daver2
September 30th, 2014, 12:40 PM
I think you have actually done the more brave thing by applying power to it in the first place...

So, if I understand you correctly, you would prefer not to make any modifications to the unit until you know precisely what the problem(s) is/are and then you can make an informed decision about how to proceed. You would also prefer not to perform any tests that may endanger the unit going forward.

In this case I would make the assumption that the power supply is OK and follow the "Testing and calibration" procedure outlined on page 9 of http://www.swtpc.com/mholley/CT_1024/CT1024_Assembly.pdf.

First though, can you identify exactly what cards you have in your unit. The procedure should be to remove all of the add-on cards from the main board (cursor card, serial and memory) and just take the unit back to the main board. I would also disconnect the connectors with all of the switches if these are plugged in (noting where each cable came from and its orientation if appropriate).

Next, I would assume that the variable resistors have drifted out of calibration. You seem to have a basic functioning unit (albeit with rubbish on the screen) but to get the rubbish you need both a video and synchronising pulses - so at least the oscillators and dividers are working (although we can't prove that they are working correctly at this point). A known component ageing is associated with the preset resistors (e.g. R4, R38, R6, etc.). I would first note what position they are in (e.g. mark the side of them with a pencil), use a screwdriver to rotate them fully clockwise and anticlockwise a few times to clean up the track, and set them back to where they originally were.

Applying power now should (according to the paragraph at the top of page 10) give you a display of random characters on the screen (probably ? or @). Slowly adjust R38 to get a stable picture on the screen (read the paragraph at the top of page 10 for more precise instructions). If this doesn't work - STOP. If it does work - adjust R4 and R6 as described in paragraph 2 on page 10.

If things still work, the next paragraph (number 3 on page 10) describes adding the memory board back into the system. Apply power. Paragraph 4 on page 10 describes how to check the 'home-up', EOL and EOF keys (by jumpering connector J3).

Paragraph 5 on page 10 describes how to perform basic tests of the cursor keys (again by jumpering connector J3).

See how far you get with this procedure...

Note that the correct power-on procedure is described in paragraph 3 on page 11 as select the desired page and depress 'home-up' first followed by EOF. You have to perform this procedure on each page separately.

If something doesn't work as expected. STOP at that point and report back.

To diagnose a fault any further will ideally require an oscilloscope. You have previously said that you own one - just not how to use it. This is probably the best time to learn! So, what make and model do you have and I can guide you through the process (first to check the power supplies for ripple to ascertain that the capacitors are OK, and then to probe into the logic). With all of the add-on boards removed, we have reduced the logic we have to debug.

Hope this is helping?

Dave

falter
September 30th, 2014, 09:03 PM
Thanks Dave.

It appears to have all the option boards available -- cursor, memory, screen read and serial. Removing them all and powering up gives me a field of @s for the most part but also some other gibberish. Looks like this:

20723

That is the best R38 can pull off. If I play with the horizontal hold control on the monitor (it's a Commodore unit), it can squish it into one thing in the center of the screen like you've seen in the video. But that pic is with everything out, trimmer adjusted to best it'll get to.

daver2
October 1st, 2014, 12:09 AM
That's good. We have taken a load of stuff (the boards that you have removed) out of the equation for the time being and it looks like the majority of th remaining logic is working - just not quite correctly. You seem to have characters present (which implies the character generator is working), and visible rows and columns of characters are definately present. We just seem to have either a synch issue or a power supply problem - we now need to identify which...

There is no guarantee that each character cell will contain the same value (e.g. '@') as the data lines from the memory to the character generator are floating (i.e. there is a high probability that they will float 'high' - but (because they are floating) they could be picking up some noise and get read as a 'low' when the time comes to display the row of dots forming the specific character. Also, there is no guarantee that successive lines will read the same bit pattern - so you could have the top half of one character and the bottom half of another! The only way to guarantee that a consistent set of characters are displayed would be to attach 10 kOhm pull-up resistors to +5V from the memory data out pins to the character generator data in pins - although I think at this stage that would probably be overkill).

You didn't specify if you are using the 'original' power supply or not. Can you please specify.

We need to start using the oscilloscope. Can you please specify the manufacturer, make and model so I can download the relevant user manual.

Cheers,

Dave

falter
October 1st, 2014, 06:58 AM
That's good. We have taken a load of stuff (the boards that you have removed) out of the equation for the time being and it looks like the majority of th remaining logic is working - just not quite correctly. You seem to have characters present (which implies the character generator is working), and visible rows and columns of characters are definately present. We just seem to have either a synch issue or a power supply problem - we now need to identify which...

There is no guarantee that each character cell will contain the same value (e.g. '@') as the data lines from the memory to the character generator are floating (i.e. there is a high probability that they will float 'high' - but (because they are floating) they could be picking up some noise and get read as a 'low' when the time comes to display the row of dots forming the specific character. Also, there is no guarantee that successive lines will read the same bit pattern - so you could have the top half of one character and the bottom half of another! The only way to guarantee that a consistent set of characters are displayed would be to attach 10 kOhm pull-up resistors to +5V from the memory data out pins to the character generator data in pins - although I think at this stage that would probably be overkill).

You didn't specify if you are using the 'original' power supply or not. Can you please specify.

We need to start using the oscilloscope. Can you please specify the manufacturer, make and model so I can download the relevant user manual.

Cheers,

Dave

I don't have an oscilloscope.. I was planning to buy one at one point but life intervened. I wasn't really sure what to buy (not knowing anything about them) so any advice on models would be appreciated. The original SWTPC power supply is what is installed at the moment and what I have been using.

daver2
October 1st, 2014, 08:06 AM
Sorry - misread your post #40 last night regarding the oscilloscope (stayed up too late again obviously)!

Looking again at the photograph you posted in #47 - is it possible that there is an incompatability in the synchronising signal levels between the CT-1024 and the monitor you are using? I am just wondering if the black bar in the middle of the screen is actually the frame flyback period? Does the monitor have a switchable synchronising feature or not?

I would be personally thinking about making up some electrolytics (with flying leads) to temporarily connect on the power supply rails just to rule out the smoothing capacitors in the power supply unit. Buy three reasonably cheap electrolytic capacitors from the local electronics shop with a similar value and a voltage rating at least as high as those in the power supply unit (electrolytics only) and temporarily connect them in parallel with the capacitors that are already there. If the existing capacitors are 'dry' and not performing their job - this should solve the problem. The use of clips (if you can find a place to hook them) will mean that you are not permanently modifying the PSU. If this fixes the problem - you can then decide about the best way of fixing it properly.

Note: watch the polarity of the capacitors when you connect them.

I will have another think tonight about the best way of continuing without an oscilloscope.

Dave

falter
October 1st, 2014, 08:33 AM
Thanks Dave. Your help is invaluable and appreciated. Looking at the back of the power supply it will be tricky to attach caps to the solder points on the backside. So as long as I'm higher than spec I'm ok? Like if I could only find 10v 8000uf rather than 10v 5000uf?

Regarding the monitor, I have a vintage Sanyo VM-4509 here, and per some instructions I found elsewhere I set it to 75ohm with the little switch in back. But exact same result. I have been wondering if we might have a video incompatibility. Reading the instructions, it seems like they were expecting a specific kind of video monitor or a modified TV set. The original connector on the end of the video cable was a small earphone jack like device, which I changed to a male RCA plug.

Also I committed a stupid last night. I was looking at a sn7422n IC that is placed on a small brown PCB perpendicular to the mainboard. I took a photo before I took it out (was just inspecting the legs which looked a bit browned/corroded at first blush)... but the camera goofed and now I'm not sure which way to put it back in. Here's some pics:

20726
20727
20728

Looking at the socket closely, it appears there is a triangular arrow on the left pointing left.. thinking that's the pin 1 directional. But I don't know to trust that given this is all home brew. Based on IC schematics I tried to hunt for 5v with my MM at all four corners but no luck.

falter
October 1st, 2014, 08:35 AM
Note: Photos came out upside down.. arrow is pointing right in the photos but left when facing right side up.

falter
October 1st, 2014, 08:40 AM
Here's the backside of that board. I'm guessing the red wire is supposed to be VCC, and black is GND, which lines up with the datasheet.

20729

daver2
October 1st, 2014, 09:54 AM
Yes, using a larger value capacitance should be OK if it is not too far out and the function is a storage capacitor (as it is). I would just go for the +5V rail capacitor first.

Thinking on the way home from work - what I said about the polarity of the synchronising pulse(s) is probably not relevant - so ignore it. Although something that did pop into my mind was exactly what you mention - the CT-1024 was designed to plug into a television and not a monitor - so the monitor that you use will have to be able to lock onto the frame/line rate generated by the CT-1024. If not, you will get strange results (especially if it is an analogue monitor - as it is).

What is that little card doing?! This looks like a home modification for some reason. The SN7422N is a dual 4-input positive NAND gate with open collector outputs. This is not part of the original schematic. Pin 14 is +5V and pin 7 is 0V/GND for this IC. Pin 1 of the IC is usually denoted by a small circular indentation or it is to the left of the cutout (with pin 14 being to the right of the cutout). Only half of the chip appears to be used. Pins 1,2,4 and 5 are the INPUTS with pin 6 being the output. Can you chase where the wires go to?

5V should be on the IC! As I said, pin 7 should be ground and pin 14 +5V so your meter should read properly.

It is interesting you mentioning about the dirty IC legs. I had an Apple IIe disk controller card which did not work. When I tried to read the ROM code back from the card - I kept getting repeating patterns that implied faulty addressing. When I removed the chips with the black legs (one of them was the ROM I was reading and the other was a 7474 dual flip flop) some of the IC pins got left behind in the sockets! The corrosion (or whatever it was) had eaten right through the thin part of the pin with a result that the IC pin was not in contact with the rest of the circuit. Replacing the components cured the problem (after de-pinning the sockets of the detached body parts)! It may be worth removing the ICs and checking that the pins are OK - cleaning as necessary.

Dave

daver2
October 1st, 2014, 09:57 AM
Sorry - brain dead!

Just realised your chips are soldered into the main board so I wouldn't recommend removing them! Inspection with a magnifying glass and a bright light called for...

Dave

Chuck(G)
October 1st, 2014, 10:15 AM
My guess on the 7422 is that it provides some support for software cursor. When I got hold of a very early version of WordStar, I realized that cursor control, however rudimentary was necessary for its function. Fortunately, a few NAND gates can provide that (I did mine with 1N914 diodes).

falter
October 1st, 2014, 08:38 PM
Here are some pics if it helps any:

20740
20741
20742

Yeah desoldering I don't think is on the menu here. All of these chips date 1975 or earlier.. I feel like I'd be trying to repaint a Ming vase. I hope our problem is just power supply related and that maybe I can create something that subs in power when we need it. Otherwise, it may be stuck like this. I just can't bring myself to mess with something this old and original too much.

Chuck(G)
October 1st, 2014, 09:10 PM
My guess is that you can remove the 7422 from its socket if you suspect that it may be causing the problem. It shouldn't affect regular operation.

Do find yourself a logic probe, however.

daver2
October 2nd, 2014, 12:31 AM
The green wire (the output of the add-on chip) goes to R23 which is the pull-up resistor for "cursor back" - so Chuck's assertion that it may be something to do with cursor manipulation appears to be correct. I can't see from the photographs where the purple and yellow wires (inputs to the 7422) are connected to - yuo will have to take a very close-in view of each connection and the components in the immediate vacinity. Assuming that no PCB tracks have been cut - then I agree with Chuck that leaving the SN7422 component out should not cause any problems.

What I have been working on last night are some simple tests that you can perform with nothing more than a multimeter, a few 10 kOhm resistors and jumper wires. These tests will (hopefully) prove what is working and - as a result - we may be able to discount bits of logic that it can't be. What we have left is then logic where the fault is likely to be (a bit like Sherlock Holmes - "once you have discounted the the impossible, then whatever remains, however improbable, must be the truth").

I will post these tests shortly.

Dave

daver2
October 2nd, 2014, 06:24 AM
I have found some evidence that the additional SN7442 chip is the 'backspace mod'. Have a look at http://www.swtpc.com/mholley/CT_1024/BackSpace/BackSpace.htm and the associated schematic diagram. The adapter card as shown is obviously more complex than your mod. as it handles CR, LF and BACKSPACE. Because your GREEN wire from the SN7442 is connected to R23 (cursor back), gut feeling is this is your mod.

Dave

daver2
October 2nd, 2014, 06:35 AM
I think I 'may' have found a video incompatability that may make the CT-1024 not work with a 'bog standard' monitor - although I need a bit of time to think through the implications.

The original Radio Electronics construction articles can be found at http://www.swtpc.com/mholley/RadioElectronics/RE_CT_1024.htm in particualr the February edition http://www.swtpc.com/mholley/RadioElectronics/Feb1975/RE_Feb_1975_pg30.jpg identifying the video output specifications. The article states that the video output pulse is 2.25V with a 1V sync pulse. This is not what I would expect for a video signal. I am used to 1V with a 0.3V sync pulse. There also may be some issues with the number of lines (frame rate) being different to a 'standard' monitor (whatever one of those is...). Again I need some time to think about this.

The voltage levels can be fixed very simply with a small plug-in adapter board on the video output of the CT-1024. The number of lines issue I need to think about...

Just some thoughts in passing based upon what I have read today.

Dave

daver2
October 2nd, 2014, 07:03 AM
On the assumption that it is the video signal to the monitor that is the problem; we should be able to 'prove' this by trying various tests to check that the CT-1024 functionality is as expected (albeit it may be difficult to see the results on the monitor). This should give us confidence that a large amount of the CT-1024 logic is actually working and we just need to look at getting a stable and readable picture on a monitor!

I am assuming that all the switches, push-buttons and expansion cards (memory, URT, CURSOR control and screen read) have been removed and all we have is the main board to work with.

Download the data sheet for the 2513 character generator from http://amigan.yatho.com/2513.pdf.

You can see on page 7-46 what characters should appear commesurate with the settings an address lines A4 through A8 (corresponding to 2513 pins 17 through 22 and CT-1024 connector J8 pins 11 through 6).

Because a memory board is not plugged in, the data lines from the memory board (address lines A4 through A8 to the 2513 character generator) are floating. This state appears to generate an '@' character which corresponds to A4..A8 being at a logic '0' level (although why not all '1's giving a '?' I do not know at the moment - possibly to do with the internal circuitry of the 2513 character generator?)

By temporarily joining each address line in turn to +5V and then 0V via a 10 kOhm resistor - we should see the character that is displayed on the screen change.

First of all connect J8 pin 11 to +5V via a 10 kOhm resistor. This should correspond to the character for 000001 or 'A'. Disconnect J8 pin 11 from +5V and connect it to 0V via the 10 kOhm resistor. The displayed character should once again return to 000000 or '@'. Disconnect J8 pin 11 totally.

Repeat the above paragraph for the following pins in turn:

J8 pin 10 - 000010 character displayed should be 'B' when the pin is connected to +5V via the 10 kOhm resistor.
J8 pin 9 - 000100 character displayed should be 'D' when the pin is connected to +5V via the 10 kOhm resistor.
J8 pin 8 - 001000 character displayed should be 'H' when the pin is connected to +5V via the 10 kOhm resistor.
J8 pin 7 - 010000 character displayed should be 'P' when the pin is connected to +5V via the 10 kOhm resistor.
J8 pin 6 - 100000 character displayed should be ' ' (i.e. the blank or space character) when the pin is connected to +5V via the 10 kOhm resistor.

If this works out - then we have pretty much proved the character generator, parallel to serial converter logic and part of the divider chain.

Dave

daver2
October 2nd, 2014, 08:25 AM
In the above post I am using the 10 kOhm resistor to limit any fault currents should you accidentally touch the wire onto the wrong place to prevent damage to your CT-1024. Thinking about the value a bit, 10 kOhms may be a bit high. 2.2 kOhms may be a bit better...

You should also be able to work out that by connecting multiple J8 pins 11 to 6 to +5V or 0V will give you a 6-bit binary pattern that will correspond to the display of a specific character identified within the 2513 character generator data sheet.

Assuming all is well to this point - I would propose that we next check out some of your switches.

I assume that your J10 has the Page 1, Page 2 and solid cursor switches connected to J10 pins 1, 2 and 6 respectively?

Connect the black (negative) lead of your multimeter to the 0V/GND of the power supply and set the multimeter to a 10V d.c. range (or whatever is convenient to read 5V).

With 'PAGE 1' selected; the voltages on IC27 pins 13 and 10 should be 0V and +5V respectively.

With 'PAGE 2' selected; the voltages on IC27 pins 13 and 10 should be +5V and 0V respectively.

If your PAGE 1/PAGE 2 switch has a centre position; the voltages on IC27 pins 13 and 10 should be +5V and +5V respectively.

Set the PAGE switch to 'PAGE 1' when you have finished with the above tests.

With the 'SOLID CURSOR' switch OFF (i.e. flashing cursor); IC8 pin 4 should be +5V.

With the 'SOLID CURSOR' switch ON (i.e. a block cursor); IC8 pin 4 should be 0V.

Set the 'SOLID CURSOR' switch to OFF when you have finished with the above tests.

Turn the power supply OFF to the CT-1024. Wait a bit and then put the memory board back in. Turn the CT-1024 back on.

You should see 'garbage' on the screen as the RAM has powered up with semi-random contents.

Temporarily pulse J3 pin 1 to ground (this should home the cursor).

Temporarily pulse J3 pin 8 to ground (this should clear the screen and leave a flashing cursor in the upper-left of the screen). Of course the definition of a flashing cursor is a bit suspect if the monitor is not synchronised properly!

Turn the 'SOLID CURSOR' switch ON. The flashing cursor should stop flashing and become a block cursor (or in your case, the lines that were flashing should stop flashing and be lit permanently).

Turn the 'SOLID CURSOR' switch OFF. The flashing cursor should appear again.

Next, select PAGE 2. The screen should display random 'garbage' characters again as the above 'clear to end of frame' has only operated on the visible page at the time (i.e. half of the memory contents).

Temporarily pulse J3 pin 1 to ground (this should home the cursor).

Temporarily pulse J3 pin 8 to ground (this should clear the screen and leave a flashing cursor in the upper-left of the screen once again).

If this works OK then we have checked out the basic operation of the memory and the cursor comparator logic.

The next installment will perform some cursor moves - although I will let you do some testing first and see what you come up with. It is pointless to go any further unless the above checks out. In the meantime, I will think about the video voltage levels and timing a bit more...

Dave

daver2
October 2nd, 2014, 08:27 AM
I got a bit worried there - the browser crashed and I thought I had lost the whole post :-(

Auto-save to the rescue :-)

Dave

falter
October 2nd, 2014, 08:40 AM
Thanks for this Dave... I don't know if I have any spare resistors kicking around.. but should be easy to find. To be honest, I have not seen any evidence of a cursor with boards in or out. It could be up in that gibberish area above or below. Perhaps it's better to go after the video so we can see everything properly? Would that adapter be hard to create?

daver2
October 2nd, 2014, 09:00 AM
A trip down to Radio Shack sounds on the cards... (We have Maplins in the UK).

If it is a voltage level issue - then no, the adapter will be dead simple (two resistors - and possibly a zener diode - at the most). You may even be able to use a 250 Ohm linear potentiometer (the same trip to Radio Shack?). Wire one end of the potentiometer to the video output from the CT-1024 and the other end of the potentiometer to ground. The output signal to the monitor is then taken from Ground and the potentiometer's wiper. The closer the potentiometer's wiper is to the video output wire from the CT-1024 then the higher the video voltage will be to the monitor. You could try this as a simple test if you wanted - you won't damage anything as it is unpowered. As you turn the potentiometer towards the 'ground' end - the 'whites' on the screen should start to dim and eventualy disappear. You may need to re-tweak R38 again...

If, however, the problem is both voltage level and timing related - then the above will not work (we are only fixing one of the problems not both of them).

The reason I am suggesting trying the tests in the previous post is to give us both confidence that the CT-1024 is at least working and that there are not loads of simultaneous problems that we are trying to chase down at the same time (which will juse send us round in circles).

The advantage of performing the tests that I have suggested above is that it should give us a clear screen with a flashing cursor (albeit potentially screwed up). If you have lots of random characters on the screen then you are correct that we won't be able to discern the garbage from the cursor. However, if the screen does not clear when it should - then there is something else faulty which may be preventing the video display from appearing correctly anyhow. We are also following the original "testing and calibration" procedure to some extent (with a few extra tests of my own thrown in for good measure).

Besides - I am now going out for an evening Curry and it is still early in Vancouver!

Dave

daver2
October 3rd, 2014, 06:59 AM
I have been unable to track down any video Standard that has the claimed voltage levels that the CT-1024 outputs. The oldest/closest I can find (for the US NTSC market that is) is RS-170 with a +1.0V (white), +0.075V (black), 0.0 (blank), -0.4V (sync) = 1.4V pk-pk. This is later reduced to +1.0V pk-pk.

From a timing perspective - it would appear as though the CT-1024 outputs an almost NTSC compatable signal with the expectation for an interlaced monitor (i.e. half of the lines are output in one frame with the other (identical) half in the next frame). If your monitor is non-interlaced then all bets are off.

What manufacturer, make and model of monitor are you using?

The monitor shown in the photographs accompanying the CT-1024 are (as I have read) a Sanyo VM4209 (although I have been unable to locate any technical details for this monitor on the web).

Based on what I have read today - trying to match the video voltage output by the CT-1024 to the specific details of the monitor you are using should work - providing the monitor is interlaced. If the monitor is expecting a non-interlaced signal - then it may be expecting all 525 (ish) lines to be displayed in one frame (which is not the case with the CT-1024).

Dave

falter
October 3rd, 2014, 06:32 PM
Right now I'm linked up to a Commodore 1702. I do have that Panasonic vm4509 but no difference in display there. Also tried my Apple /// monitor. What is throwing me is the original video connector.. like I said it was one of those small earphone style numbers. Ive never seen an input on any tv or monitor like that.

I did try some experimentation. The machine definitely reacts to the switches. On one occasion I fired it up with only the cursor board in and actually briefly got a blinking cursor for two seconds, about five lines down in the field of gibberish. Ibalso noticed that when I pressed the EOF button a number of the gibberish characters would change.

With the mem board in, I get a single illegible line of gibberish where the top of the screen should be. There is, if you look closely, a reaction when you press certain buttons. But it returns to same as before when released. Keyboard and shorting keyboard pins produces nothing.

I kind of wonder about bad ram.. is it safe or possible to do the piggyback test we used with my PET?



I have been unable to track down any video Standard that has the claimed voltage levels that the CT-1024 outputs. The oldest/closest I can find (for the US NTSC market that is) is RS-170 with a +1.0V (white), +0.075V (black), 0.0 (blank), -0.4V (sync) = 1.4V pk-pk. This is later reduced to +1.0V pk-pk.

From a timing perspective - it would appear as though the CT-1024 outputs an almost NTSC compatable signal with the expectation for an interlaced monitor (i.e. half of the lines are output in one frame with the other (identical) half in the next frame). If your monitor is non-interlaced then all bets are off.

What manufacturer, make and model of monitor are you using?

The monitor shown in the photographs accompanying the CT-1024 are (as I have read) a Sanyo VM4209 (although I have been unable to locate any technical details for this monitor on the web).

Based on what I have read today - trying to match the video voltage output by the CT-1024 to the specific details of the monitor you are using should work - providing the monitor is interlaced. If the monitor is expecting a non-interlaced signal - then it may be expecting all 525 (ish) lines to be displayed in one frame (which is not the case with the CT-1024).

Dave

daver2
October 4th, 2014, 04:35 AM
I recon the Commodore 1702 should work (assuming connected to the front-panel composite video connector). You have tried adjusting the vertical hold control on the monitor haven't you?

I don't think the connector is relevant. The CT-1024 was designed (in part) to be connected to a modified television - so the constructor could have used any old type of connector they liked.

Randomly pressing buttons with random cards inserted doesn't help diagnose the problems. Only a logical progression will.

If you think about it - it can't be a RAM fault if we have the RAM board (and all of the other optional cards) removed. My hypothesis is that if the main board doesn't work on its own - it can't have anything to do with the add-on cards as they have all been removed. The problem(s) must lie with either the power supply and/or the main board.

If you can't get a stable picture on the screen without the option boards present - the fault must lie somewhere with the timing chain/counters and/or synchronising logic. Once we have solved the problem with those - we can move on to the next problem(s) in sequence. We can only fault find on the RAM board when we know for a fact that the main board is solid.

My advice is to add the smoothing electrolytics as we have discussed and to perform the relatively simple tests I have already suggested to see what happens.

I have now read up on how the vertical and horizontal synchronising pulses are generated - but there is no point looking into this until we have some answers to the basic operation (i.e. are some of the basics working at all).

I noticed some earlier posts you made on YouTube. It is interesting to see that some of the characters appear to be "well formed" - but there is a lot of what I would call "intermittent operation" - may be from bad solder joints, possibly suspect capacitors or power supply problems. But this is me jumping to conclusions...

Dave

billdeg
October 4th, 2014, 08:17 PM
early 1975 poptronics mag stuff