Jump to content

apm

6502
  • Content Count

    180
  • Joined

  • Last visited

3 Followers

Recent Profile Visitors

389 profile views
  1. I replaced the TDA4605 and recapped the whole board. Not sure if the results are progress. Now with the video cable unplugged, the monitor powers up and makes an ominous sizzly noise. I think it's coming from the flyback but I'm disinclined to get close enough to really pinpoint the source. I may have heard a couple arcs, but I can't see anything even with the lights off. There's a dim raster that overscans the CRT. With the video cable plugged into a Mac which is turned off (i.e. no signal), the sizzling stops and I get a bright white raster, wider than the screen and a bit too short [1]. With a video signal present, it goes back to power cycling like it did before. So where's the sizzling coming from and why? Damaged flyback perhaps, or maybe damaged sweep IC (TEA2037)? Perhaps even an ongoing power supply problem: I don't have a meter or scope to hand right now, so I can't check the voltages. Perhaps they are still too high. As it happens, I was able to do some testing on a working unit recently. The normal power supply voltages are around 40V and 12V (mine would rise to 50V and 18V before shutting down). The mystery RP13 is a vertical ceramic resistor of some sort -- see picture -- which is unlabelled but (from memory) measured between 5k and 6k. [1] That plugging in a cable, even without signal, changes things is not unexpected from what I understand of the schematic. There are a few odd bits of circuitry on the analog board which take their input from one of the monitor sense pins that's grounded when you plug it into a Mac. One bit of circuitry looks like it would increase the voltage to the flyback primary by maybe 2V when a Mac is connected. Another looks like it would add a 1200pF capacitor in parallel between the drain of the HOT and ground. I'm not sure what purpose either of these circuits serves or why they were designed in.
  2. Good thoughts, thanks. I think the HOT and its driver (TEA2037) are okay: there's a sensible signal at the HOT gate (not itself diagnostic), and a reasonable voltage appears on the one flyback secondary I felt comfortable to test with a 10x probe. When I retrieve my 100x probe from the office I can probe the higher-voltage parts of the horizontal circuit, but it looks like that's not the source of the problem. I will have to get my hands on an isolation transformer. Any particular reason to put the isolation transformer on the scope rather than the monitor? I guess I could also just try replacing the TDA4605 and see what happens.
  3. DP5 is on the primary side of power supply. Why it shorted I'm not sure. See attached reverse-engineered schematic. (Apologies for low-quality photo; no time to format this properly right now.) A few components I couldn't read the value, especially the SMT caps. In any case, I believe RP13 serves as a kickstart to the TDA4605 by providing a trickle of current from the unrectified mains line. It looks like after the circuit starts up, the winding of ZP1 pins 3-4 would provide a source of continuing power to the TDA4605. In a circuit in the datasheet for the TDA4605, a similar resistor to RP13 has a value of 75k. I tried 47k and then 22k, which is what's in the later Classic analog board. Same result in each case. I don't think it could go much lower without needing a >2W part, which doesn't seem likely given the size of the pads. I can't put a scope on the primary side of the circuit, but I did scope the secondary side tonight. There are two secondary voltages, a higher one used for horizontal sweep and the CRT neck board, and a lower one powering most of the rest of the circuits. See attached. Both voltages rise together until they reach around 50V and 18V, at which point the regulator shuts off, and the cycle begins again 500ms later. I don't know what the voltages ought to be when stable, but I'm wondering if this is an over-voltage crowbar circuit kicking in. In the brief period these two voltages are active, the horizontal sweep runs: I can see a periodic signal on the gate of the HOT, and voltages appear on (at least) one of the flyback secondary windings.
  4. I've worked out a schematic for quite a bit of this monitor by now. It has some similarities for the later Classic analog board, though with considerably more complexiy in the video circuit for various types of geometry correction. The CRT neck board looks like a pretty good design on which to base a grayscale adapter for compact Mac CRTs. More on that later. In the meantime, replacing DP5 (an 18V zener which was shorted) and trying 22k or 47k for RP13 results in a "flup-flup-flup" power cycling where the LED transiently comes on and then the system powers down, repeating the cycle about 2 times per second. It's possible a different RP13 is needed (anyone have one of these monitors??). But I'm wondering what else might cause this problem.
  5. apm

    Yet another SoC to B&W CRT thread

    The main challenge with VGA on the internal monitor is that the horizontal scan rate is different, and this requires substantial modification to the existing analog board. There's a secondary challenge around grayscale vs. black and white, which needs a different CRT neck board, but that's an easier challenge and also less of a gating factor to getting something the screen. I started down the resolution-changing road a few years ago, and I'm still (very) slowly working on it in whatever bits of free time I find: The good news for you is that you've got a Classic II rather than any of the earlier compact Macs. As you'll see near the end of the above thread, the Classic II (and late-model Classic) have a different analog board design with a different flyback transformer than all the previous compacts. It appears that this flyback is amenable to being run at a higher scan rate, where the earlier one basically can't go above 22kHz. The starting point would be to decrease the value of the flyback capacitor, then generate an HSYNC signal with the right pulse width to replace the HSYNC coming from the logic board. You will probably need some external circuitry for that (e.g. a one-shot timer). You may also need to find a way to generate a higher power supply voltage for the horizontal section to get a wide enough picture, but I haven't gotten that far yet. Certainly the 9" VGA monitor option is easier if you're taking out most of the guts, but I do think VGA will eventually be possible with the built-in analog board, at least on the Classic II.
  6. apm

    Mac SE/30 with MicroTouch touchscreen

    I was thinking the same. Is this an internal card or an ADB device? snuci do you have pictures of the board(s)?
  7. apm

    Macintosh SE/30 with Radius Card

    The internal monitor will work fine with the Radius card. The black screen will be due to a fault with either the logic board or the analog board. If it's an SE/30, start by recapping the logic board. The Compact Mac forum has a number of threads that might be useful for how to diagnose having no picture.
  8. I'm working on repairing a non-functioning Macintosh 12" Monochrome Display. The proximate cause seems to be that resistor RP13 is missing -- completely gone, nowhere to be found, nothing but a blackened set of pads left behind. The underlying fault appears to be that DP5 has shorted, but replacing it doesn't do me much good until I figure out what to replace RP13 with. Does anyone have one of these displays who could open it and post a photo of the main PCB? RP13 is just behind the mains socket, though I can't even tell if it was originally a through-hole or surface mount part.
  9. apm

    Not a "Classic" Checkerboard Issue

    The checkerboard is from caps on the logic board, but the unstable image is almost certainly the analog board. It will also need a full recap. I've come across all sorts of unstable screen issues from bad caps in the Classic analog boards. For whatever reason they are often in much worse shape than the SE(/30) and earlier models.
  10. apm

    Feeding VGA 31KHz to a Classic II CRT?

    For what it's worth, with a Classic II the stuff I was trying to do on the "Compact Mac retina display" thread might actually be possible. The flyback used in all the compacts up through the early Classic basically can't be run much above 22kHz -- there's so much internal capacitance in the windings that you can't get a short enough flyback pulse to run at higher horizontal scan rates. The flyback in the late-model Classic and the Classic II looks more flexible. See the end of the thread. I haven't had much time to look at it in the past several months but I still plan to revisit it when I get a chance.
  11. It's definitely a vertical sweep problem, so as @techknight says, nothing to do with the flyback. What you don't know is whether the problem is on the analog board (vertical drive circuitry) or the logic board (no vertical sync signal). A scope is the right way to test that, but if you have a multimeter with a frequency measure, put it on J4 pin 5. You should get a 60Hz signal. Given the other voltages it does sound like the problem might be in the power supply. U1 (op-amp LM324) is shared between power supply and vertical circuits. It's also a cheap and common part, easy to replace. Since both vertical and power supply circuits seem amiss, you might try swapping it.
  12. apm

    Vertical bars on an SE

    Never mind, you're right. I had remembered that the SE had a faster video circuit compared to the Plus, but they both use main memory. It's hard to debug this without a schematic, but the closest I can find is Figure 12-1 (p. 251) in Designing Cards and Drivers for the Macintosh Family. Video on the SE is generated by the BBU, the big square custom chip on the logic board. It grabs 16 bits of video at a time out of main RAM which seems significant in light of the 16-pixel-wide bars. A broken trace between RAM and the BBU is one possible culprit, more likely an address trace than a data one, though I'd think that would produce a duplicated image rather than white stripes. Or the BBU itself could be faulty. It's socketed so a swap could test that quickly. Or there could be something else entirely going on that's preventing the BBU from accessing RAM in time.
  13. apm

    Vertical bars on an SE

    If it was a RAM issue you'd have a Sad Mac, or at least other crashes. I'm pretty sure that the SE, like the SE/30 but unlike the Plus, has separate video memory that doesn't use main RAM. Unfortunately, I can't find any schematics of the SE video circuits. The logic board schematics online are missing that page. Anyone around here have them? The problem will be somewhere in the video circuit. An interesting test can be to take a screen shot (Cmd-Shift-3) and look at the result. If the screen shot shows the jail bars, then the problem is getting the data into the video memory -- probably a broken trace on the address lines somewhere. If the screen shot is clean, the problem is between video memory and the screen.
  14. Here's a weird idea, maybe not a good one but probably not harmful (@techknight?). Try shorting out R29, a 22-ohm SMD resistor on the bottom side of the board, just under UI6 (I think). R29 is in series with the CLK16M signal (i.e. the main system clock), which is generated by UI6. If there's enough capacitance on the bus, for example through too many cards, then the rise time of the signal could be just barely too slow. But shorting that resistor will make the drive just a little bit stronger, which might be enough. I imagine it is there to meet emissions specs and that shorting it probably wouldn't damage anything, though I suppose there's a chance that it could create a stability issue. The even stronger version of this would be to place a 5V high-speed buffer in place of R29. However I still don't have a good explanation for why a laggy clock signal would produce this pattern on the display, so maybe it won't make any difference.
  15. Interesting! The question then is whether UE8 is faulty, or if it's just sitting on the very edge of spec because of the other load on the PDS slot. I wouldn't bother heatsinking UE8: try replacing it and see if the problems still persist.
×