On 08/14/2016 08:04 PM, Fred Cisin wrote: Didn't RCA make computers? Don't know who VGA was, though. Vertical Gizmo Authority?

I assume the 'CGA' in the name in the silkscreen is relevant! It only has 16K RAM, a pair of 4416s so it's not EGA or VGA. BUT.... There were a lot of 3rd party cards that could be configured (DIP switches) to be either MDA (often hercules compatible graphics) or CGA. You had to use the right monitor, of course. A trace going to pin 7 of the DE9 is a clue (it's unused on the original CGA board, it's video on the MDA) but not always. Some CGA-only cards put the composite signal on pin 7. -tony

That one has a lightpen connector, which his dorsn't. Not usually important, but it might have other differences, such as different functions for the switchrs. There were some MDA boards without printer. Plugging 5151 into CGA could damage the 5151. Even plugging it into EGA with wrong switch settings could. On Sun, 14 Aug 2016, Chuck Guzis wrote:

This is a CGA clone. It can work with CGA and Hercules monitors/modes, and usually were switched by a program called MONO.COM. I Believe I have some boards like that. Possibly I have the manual and disk. 2016-08-15 1:53 GMT-03:00 Chuck Guzis <cclist at sydex.com>:

That depends on the MDA/hercules monitor. The IBM5151 does not have a horizontal oscillator. It will attempt to run at whatever frequency you give it. Give it the wrong horizontal frequency and the horizontak output stage will not like it very much. Most other MDA type monitors were more conventional with an intenral horizontal oscillator that attempts to lock to the incoming sync signal. In any reasonable one, if the incoming frequency in way off, the oscilllator will not lock to it, it will remain running at a frequncy close enough to normal that there will be no damage. The common Zenith MDA monitor, the one with that well-known power supply [1] is like that, [1] The one that combines the efficiency of a linear with the reliability of a switcher. And which malfunctions if you replace the green power-on LED with a red one. Ask me if you have seriously never come across this. In any case it rarely costs you the monitor if you get it wrong. A horizontal output transistor, sure. But they are not that expensive. -tony

I am not joking. This PSU is a 'classic'. Let me describe the circuit. The monitor circuitry needs 12V DC. This powers the CRT heater, the small-signal circuits and the horizontal side of things. In the conventional way, other voltages are produced from the flyback transformer. So the PSU has to provide 12V DC from the mains. Here's how it does it. Mains comes in and is rectified/voltage doubled to get about 350V DC. Just like in a switcher. This feeds a chopper transsitor/transformer circuit. But there is NO regulation applied to that chopper. It free-runs, using an extra winding on the transformer to drive the base of said transistor. This circuit produces about 18V AC. It _is_ a switcher at this point, with the advantage that the transformer and secondary-side smoothing capacitors can be smaller than they would be if it was a mains-frequency circuit. On the other hand that chopper does fail! OK, we now have 18V or so AC from the chopper transformer. It is half-wave rectified and smoothed. It now feeds a _linear_ regulator circuit made of discrete transistors. Hence my comment about the efficiency of a linear. It is a linear PSU now. But there is another quirk up its sleeve. A linear regulator works by comparing the output voltage to a reference voltage and controlling the pass transistor appropriately. Conventionally that reference voltage comes from a zener diode. Not here. They used the forward voltage of the power-on indicator, a green LED, for this. OK, the forward voltage of an LED is approximately constant, and doing this saves a zener diode (a few pence/cents). Of course if you happen to replace that green LED with a red one (lower forward voltage), the PSU output voltage goes down too. -tony

Of course I was being ironic. The 'not joking' part refers to the fact that it really is a switcher with a linear regulator. The reverse (mains transformer feeding low-voltage switching regulators) was relatively common (think of the older PDP11 PSUs, for example). These are big and heavy (massive iron-core transformer) and not as efficient as a mains switch-mode PSU, but are more efficient than a linear regulator (obviously) and seem quite reliable. -tony

Of course. But the PSU in that Zenith monitor only has one output -- +12V. And there is no regulation feedback loop to the chopper stage. On multi-output switchers often the main out has a much higher power limit than the others. Perhaps 5V at 60A and +/-12V at a couple of amps In which case it is not unreasonable to use linear regulators for the latter, the power wasted is not that much. As an aside, there is at least one DEC printer (LA324 I think) where the main output from the SMPSU is the 36V one (that is the one that is sensed to regulate the chopper), used for the motors, printhead, etc. The 5V rail (used for the logic, of course) comes off a linear regulator. One to watch for... You can of course do it with older technology too. The PDP11/44 PSU and the HP9845 PSU spring to mind. A lot of ICs, chopper transistors, etc in there. -tony

On Mon, 15 Aug 2016, tony duell wrote: Yes, readily repairable, and therefore "cost a monitor" is an exaggeration. But, somebody who is still unaware of the problem, and blithely plugs anything in that will fit, without knowing what kind of card they are connecting to, might not have the basic expertise to fix it. For example, at the college, they purchased replacements and dumpstered every monitor that failed. (I am referring to the local community college, but I know of examples of the same at UC Berkeley!) I am pleased to hear that aftermarket competitors made appropriate changes to protect against that vulnerability. I just looked at Wikipedia, and found some amazing errors, such as statement that IBM mono came along much later than CGA, caling it MDA instead of MDP, that Hercules innovatively had a "Centronics port" (sic) because that was much faster than the serial printing with the IBM mono, that PCs HAD TO HAVE both displays, because "no business software would work on CGA", no mention of the "Hercules clones" that Jenkins had tantrums about, etc. I edited a few of them, but there is just too much misinformation.

On Mon, 15 Aug 2016, Chuck Guzis wrote: Than you for the words of wisdom. I have difficulty resisting correcting even minor errors of fact. It is distressing to me how badly history can get mangled over what seems to me to be a very short period of time. Admittedly, we all see different parts of the elephant, and some topics, such as early history of the internet will always suffer from incompatible perspectives. But, it's hard to see how somebody could think that Comdex originated as a successor to the West Coast Computer Faire, or that IBM Mono was a later development after CGA.

On 08/14/2016 10:15 PM, Fred Cisin wrote: I think that the switches mostly relate to output configurations; that is, they don't change the basic operation. For example, the "monochrome composite" is probably just luminance, blank and sync (VBS), whereas the "color composite" is the complete CVBS signal. I suspect that setting either of the two composite modes probably disables the output at the DE-9. It would be interesting to see what other combinations of switches produce something useful. Of course, this is all wild-hare speculation on my part, as I don't have anything like this card sitting in my hellbox. FWIW, Chuck