While I was working on trying to get 80 columns on a ][+ with the Ralle Palaveev A2VGA card I ran into a problem with one of my 80 column cards. I have a Videx and another Videx clone that are working but my Franklin 80 column card (which is Videx compatible but a little different) isn't working right. For some reason the sync appears to be completely out of whack. The screen rolls diagonally like both horizontal and vertical sync are out.
I've tried replacing and/or swapping a number of chips on the board and it hasn't made a difference yet. I did order a new 6545 to try that. There are still a few of the 74LS chips I haven't tried yet too. I'm thinking though it might be a cap. In most of the places the board uses an odd type that I haven't seen before. I don't know how reliable this kind of cap is, whether it is like a typical ceramic that are pretty stable or like an electrolytic that tend to go bad over time. I need to get a schematic so I can find out what the values are for the caps. I also need to look at the data sheet for the 6545 and see how the sync signals are output, maybe tracing from there will help.
Anyone has any ideas I'd love to hear them...
Oh, I should add that a NTSC->HDMI converter just says "no signal" to the output of this card. Plugging it into a composite (Sanyo) monitor, I get the diagonal rolling display.
Your picture suggests there's a video signal of an Applesoft prompt and cursor, but no HSYNC nor VSYNC. That means signals are passing through the pixel-generating circuits but not through the SYNC circuits. Videx's schematic shows exactly 3 components that could block the SYNC circuit while allowing pixels through:
Although we don't have Franklin's schematic, the designs are sufficiently similar that it's probably practical to use a continuity meter to locate similar connections on the Franklin board. In particular, pin 39 and pin 40 of the 6845 display controller may be connected directly to two pins on the 74LS02.
Franklin 80 column card sync connections.png
Swap out the 74LS02 if you haven't tried that already. If that doesn't give a quick answer, try using a continuity meter to find where the HSYNC and VSYNC signals go. If they do indeed connect to the 74LS02, use a data sheet to identify the output pins and see if you can locate that 20K output resistor.
Thanks for the ideas. I already tried swapping out the 74LS02, but I will try another one out of a different batch just in case. The resistor you point out is indeed red-black-red. I metered it and it appears to be approx 19k. I don't have any 20k resistors in stock. I got 22k, but not 20. I think I will try ordering some 20k resistors and swap the one on the board out, maybe 19k isn't close enough, possibly the part has drifted out of spec over time.
The three transistors near the output jack are all 2N3904.
I tried putting a Hitachi 6845 clone in the card and with that it gives no picture.
Continuity between pins 39 and 40 of the 6545 and pins 2 and 3 of the 74LS02 appears to be OK.
At this point I am thinking maybe it is the 6545. I wish I had a scope to check the signals coming out there.
What are suggestions for an inexpensive scop that would be good for this kind of testing? It is something I've needed for a long time but have never bought. Back in the old days good scopes were super expensive and the ones that weren't most people said were basically useless. Nowdays I think there may be some usable ones that aren't bank breakers for a hobbiest who dabbles in hardware.
Looking on Amazon, it appears there are a bunch of "automotive" oscilloscopes that are under $100. I'm assuming those are no bueno, kind of like the relatively inexpensive ones were back in the 1980s-1990s. Then there are some around $200, some around $300 and then some around $400. I'm wondering if I could get by with the $200 range ones or whether I should "buy once, cry once" and get one of the more expensive ones. I'm thinking I will probably regret it if I buy one of the cheaper ones and then run into something I need the more expensive ones for. It looks like the $300-ish units are typically approx 100Mhz single channel, $300-ish are 100MHz dual channel and the $400-ish ones are 200Mhz dual channel. Based on that I'm thinking the $400-ish ones are more likely to work for a lot more things.
Anyone know anything about the Siglent Technologies SDS1202X-E? Is it any good or should I look into something else?
Oopsie, somewhere I goofed!
I gave the correct color-code but typed the wrong value every time. Yes it should be red-black-red, but that's 2K. Not 20K.
It's a 2K resistor functioning as a voltage-adder against the 1.5K resistor carrying the pixel signal.
Videx voltage adder.png
So, concentrating on those resistors for a moment:
So there's a credible likelihood that your 19K measurement indicates a fault in the 2K resistor. There's an easy test: clip another resistor across the 2K resistor to let the sync signal through. If bypassing that resistor causes the monitor to sync then the 2K resistor on the board in faulty.
It's OK if your test resistor isn't exactly 2K. If it's a lower value than 2K (eg: 480Ω) then this SYNC test will still be valid, but the pixel signal will be abnormally weak -- and that's okay because you're just trying to determine if the SYNC was getting blocked at the 2K resistor.
I checked that resistor is 1.9k, I had been thinking 20k, so I misinterpreted the decimal point.
I just installed one of these tonight. It appears to be NOS. It came with the box and manual. I get get the schematic if you still need it.
I did notice a difference.
You have all 4 wires from the card going to the header on the motherboard. Did you make that cable?
The instructions say to attach the yellow wire to pin 4 on the 9334 chip.
PXL_20231211_053459813.jpg
If you can find a way to scan and upload the schematic here that would be great.
The cable was on the card when I got it. I think the cable only matters for the softswitch function and shouldn't have any effect on the issue I'm having.
I think the cable that is on my card is for a Franklin ACE 1200, and I think they have the yellow wire signal on the 4 pin jack on the mono on that machine.
I tried a different 6545 on the card yesterday and no change.
The instructions also mentioned that on the Franklin ace, the yellow wire should attach to the pin header next to the 4-pin header.
That may be the wrong cable for this card, as it only has three wires for the 4 pin header, and the separate yellow wire attaches elsewhere on the Apple and Franklin.
\Whether it is the right cable or not, I think the card should work in 80 column mode without it hooked up at all. The cable just brings the 40 column video from the motherboard and provides the signal to switch between the 40 and 80 column video. I don't think it would affect the synch at all.
You'd know better than I. I was just offering a suggestion.
I will try to scan the schem soon but it is longer than my scanner bed.
Sorry, I didn't intend it to come across like that.
If you have to scan in two parts that's no problem. Just make sure there is some overlap in the middle and then it can probably be stitched back together in a graphics tool I would think.
No worries, I just meant that I am fairly new to these types of cards and I'm sure you know a lot more than me, since I read a post that you've been in the apple repair scene for quit a long time.
Going to try to scan the document tonight. Hopefully it helps you out!
It should help, and it will be great to archive the scan here in case someone else might ever need it. This card is similar to the more common Videx and direct clone cards, but has some differences like the built in 40/80 switching and the fact it uses a 6545 instead of a 6845 or the more common Hitachi clone of that chip. The 6545 is close, but has a few minor differences.
Here you are, sir!
franklin80col.jpg
Excellent!!!! Thanks!!!
It's substantially the same circuit as VideoTerm, but with an integrated soft-switch that can be activated/deactivated by choosing an odd vs even address between $CE00 and $CEFF. (ie: the memory page immediately after the 512-byte window into the display buffer.) That might seem like a weird scheme
Franklin didn't insert any components between the display controller's sync outputs and the input to the 2N3904 transistor, so it doesn't suggest any new suspects. But the schematic is nevertheless helpful for matching components by their part numbers. (I've used the same color coding as the previous schematic: blue for pixel signal path, yellow for the sync path, and green for the combined pixel-plus-sync signal)
Franklin 80 column card schematic.png
The aforementioned "yellow wire" appears to be a white wire in the OP photo. If that cable in the initial photo was attached to the aux video connector on the motherboard, then it would have been connected to the +12 volt power supply instead of the TEXT / GR signal. (Well...probably. The pins are numbered strangely, and there are plenty of ways to short-circuit the power rails.)
The intended function of J2-4 is to switch back to 40-columns whenever graphics mode is enabled. But Franklin omitted any pulldown resistor so those gate inputs would be left floating if it's not connected, thereby causing Q2 to weaken or block the signal. It's not a very useful feature anyhow, so I would suggest connecting J2-4 to ground -- either directly, or through a 1K pulldown resistor.
But although the board wouldn't reliably generate video if J2-4 was left unconnected (or connected to +12v), I sure can't guess whether that could cause the loss-of-sync like the picture at the top of the thread.
My Franklin card makes the horizontal video stretched slightly beyond the border of the screen, which makes it impossible to see the text at the prompt, at least on the LCD display I am using that doesn't seem to have any adjustments.
Is that normal?
I suppose I really should try plugging in that little cable. Actually I think it would be safer to make another one and hook it up like shown with the yellow wire going to the 9334, Or maybe I should work on getting one of my Franklins put back together completely. I think the cable that is on my card will plug directly into the newer ACE 1200 I have. Of course I'll need to verify that first.
You could try tinkering with the display controller's registers and search for settings that work better with your display.
Does Franklin document the CRTC register settings in the manual? If not, refer to Videx's Videoterm Installation & Operation manual, which defines all the CRTC registers on page 5-2 and lists the default values Videx used on page 5-7. Although the cards used different display controllers ICs, they should both use these registers for horizontal settings:
To try adjusting the registers, activate 80-column mode and use CALL-151 to go to the Monitor. From there you can write a new value into any register by entering address C0B0 followed by a colon, then the register number, a space, and finally the new register-value in hexadecimal. (The registers are write-only, so you can't read their current settings unless you disasseble the initialization routine in the ROM.)
For example, to set register R2 (Horizontal sync position) to Videx's default of $62, enter the command:
C0B0: 02 62Where
C0B0is the DEVSEL address for the card in slot 3,02chooses register R2, and62is the value to be stored in register R2.Then try entering a few different values and see how that adjusts the display position and geometry, like:
C0B0: 02 60C0B0: 02 5EC0B0: 02 64Try the other registers if that doesn't help. Maybe adjusting to register R3 (sync width) might affect the picture position or width, like:
C0B0: 03 25C0B0: 03 2CNo doubt it will require a bit of experimentation just to determine whether these registers do anything helpful!
The manual does say the yellow wire does not go in the 4 pin header, even on the ace 1200. It goes to a separate prong near the header.
https://www.applefritter.com/files/2023/12/13/franklin80.pdf
https://www.applefritter.com/files/2023/12/13/scan0001%20-%20Copy.pdfhttps://www.applefritter.com/files/2023/12/13/scan0002%20-%20Copy.pdf
OK, not sure why the cable that came with the card is like it is, but I guess maybe it is good I never tried to plug it in that way.
Thanks for the scans! That definitely helps!
A very interesting development... I happened to find another Franklin 80 column card stuffed in my ACE 1200 motherboard... And extremely interestingly, it does exactly the same thing as far as sync as the other one. So now I am thinking that maybe it does need that jumper for some reason. Seems weird that it does but it seems unlikely that two diferent cards could be broken in exactly the same way. I also tried the cards in a different ][+ clone (Micro II). I am thinking I may drag my Pineapple downstairs and see if it does the same thing.
Well, I'll be damned... I removed the "yellow" (white on mine) wire from the connector on the motherboard side and plugged it in... and it works...
franklin_80_working.jpg
I'm surprised it needs that plugged in, but it obviously does.
Now I need to diagnose another problem with one of the cards... I'm getting a few garbled characters on the screen. It looks like maybe the 2Kx8 SRAM may be bad. I think I have one of those somewhere. If not I will try to order one.
Is it always the same characters that are garbled no matter where they happen to be on the screen? If it is, the font EPROM has probably started flipping its bits.
No, it is characters at fixed spots on the screen. Most of them should be blank. It looks like bad RAM. I had swapped a bunch of 74LS chips on the card though. I will put back the originals and make sure nothing changes.
After I put back all the original 74LS chips if that doesn't change anything I will try swapping the SY2128 (6116) from the other card to see if that makes a difference. I suspect that will do it. If so, then I'll order a replacement. They're available and not stupidly expensive... $5-$10 with shipping.
If it is for some reason the EPROM I can copy it from the other card. I have 2716 on hand. Thankfully it isn't a 2708 or I'd have to make an adapter or something because I don't have any 2708 or a way to program them since even my old TL866CS won't do those. I have the adapter to allow external Vpp, but I'd have to modify the Open Source minipro utility to add code to handle a 2708 because it doesn't know about that.
Oopsie, sorry I inadvertently buried that requirement near the bottom of comment 22.
With the benefit of your discovery, it's easier to see why that weak signal doesn't sync if you don't connect the J2-4 jumper: the 80-column video and sync will leak through transistor Q1 as a varying impedance, but at too-high a voltage for the sync to be detected. (No danger, though. The elevated voltage is much too weak to cause any harm.)
Candidly, I think Franklin should have put a 10K pulldown resistor at J2-4 to enable the video-output by default, regardless of whether J2 is connected. Such a 10K pulldown resistor wouldn't affect the circuit when J2 is connected, but it would pull down the inputs to the LS10 whenever J2 was left unconnected. If I owned this card, I would insert this resistor just because it seems like a better circuit design.
Franklin add 10K resistor.png
Did you try pressing return until the display scrolls?
If it's bad memory then those "bad" characters should change screen position when you scroll the display because hardware scrolling works by reassigning memory addresses associated with the on-screen characters (CRTC registers R12 and R13) instead of moving the characters within the text memory. To see how they do that, see Videx's comments about "increment the start address by one line" in the "CRLF" routine in the ROM listing. Franklin used this same method because "hardware scrolling" was an advertised feature.
That's a novel contrast from Apple's built-in display modes, where each bad bit only affects one absolutely-fixed position on the screen.
I'd have to test again, but my card doesn't even seem to switch back to 40 columns after entering GR from 80 columns mode.
I think I had to reset to get back to a prompt.
This is normal behaviour with the Videx. I'll bet yours is similar.
I ordered a new SY2128 and it showed up... solved the garbage character problem so both of my Franklin 80 collumn cards are working perfectly now...
Nice!!!