• Please review our updated Terms and Rules here

Jameco + nec v20

I was starting to think about a card causing the problem. I don't have a memory board, or even a multifunction board. But I do have a more modern floppy controller that supports 1.44m drives. Since I'm not hearing any error beeps, I'm wondering if it is freezing at the point the controller usually searches for the drives. I'll have to try that out, but I think I might put that on the shelf cause I want to get my new xt-ide card built and running. ;)

I wonder specificially in the above case what was causing the conflict since the AST is a multifunctional card.
 
Last edited:
Well, now I'm really stumped. I tried a 10mhzt v20 that I got in the mail today. Then, I systematically removed each board one at a time. No dice. My system just doesn't seem to like these chips. I put in a CGA card so I could see if any video and error messages were being displayed. No video. Any other ideas?
 
Can you burn your own ROMs? I may have something that will help tell you what's going on.

But I thought you mentioned that the board (Jameco) had a V20 in it? I'm a little confused here.

If not, can you identify the board on the TH99 site?
 
Sorry, I shifted gears a couple of times in this thread. The work I've been doing here is on a genuine 5160 xt board. I'm trying to get a v20 chip to work in it in place of the 8088. I bought two 8mhzt's v20 chips from Jameco that don't work on it. Then the Jameco motherboard I bought of Ebay (and the one I was asking about at the start of this thread) arrived in the mail. So for good measure, I tried the v20 chip (10mhzt) that was installed in it to rule out a bad batch of 8mhzt v20's. And yes, I have a Willem programmer.
 
I just pulled out a spare 5160 motherboard (the 640KB type) from my cupboard, replaced the 8088 with an 8MHz V20 (D70108C-8), connected a speaker, set switches 5 & 6 to off, then powered on. As expected, it gave some beeps a few seconds later (BIOS could not find MGA video circuitry to initialise).
 
Do the 8 Mhz V20s work in the Jameco board you received ?

bugman2112 said:
Sorry, I shifted gears a couple of times in this thread. The work I've been doing here is on a genuine 5160 xt board. I'm trying to get a v20 chip to work in it in place of the 8088. I bought two 8mhzt's v20 chips from Jameco that don't work on it. Then the Jameco motherboard I bought of Ebay (and the one I was asking about at the start of this thread) arrived in the mail. So for good measure, I tried the v20 chip (10mhzt) that was installed in it to rule out a bad batch of 8mhzt v20's. And yes, I have a Willem programmer.
Click to expand...
 
Well after a long night of trouble shooting, I think I figured out what threw me off. The v20 doesn't initialize at a cold boot (maybe not really the cause, but thats the best way I can describe it). If I flip the power switch on after several minutes of being off, it won't boot. If I turn it on, turn it off, then toggle the switch back on within a few seconds, the computer boots normally. I do not experience this with the 8088. Is this perhaps a sign of a weak power supply with aging capacitors?
 
bugman2112 said:
Well after a long night of trouble shooting, I think I figured out what threw me off. The v20 doesn't initialize at a cold boot (maybe not really the cause, but thats the best way I can describe it). If I flip the power switch on after several minutes of being off, it won't boot. If I turn it on, turn it off, then toggle the switch back on within a few seconds, the computer boots normally. I do not experience this with the 8088. Is this perhaps a sign of a weak power supply with aging capacitors?
Click to expand...
On power-on of a computer, dynamic RAM usually has to go through a certain number of recharge cycles before it can be used (e.g. the datasheet for the MN4164 specifies 'several' cycles with 8 being adequate). So the author of the POST would I expect be initialising the timer/DMA to do refresh of RAM, and then making sure that none of the RAM is tested until an adequate delay passes. That delay could simply be other initialisation code or a dedicated delay created by code running in a finite loop.
If that POST is run faster, the delay might become small enough to become inadequate, resulting in the 'base 16K RAM' test failing. 5150 and 5160 motherboards just stop on such an error (with no indication). Delay loops running faster than they were meant to can cause other timing issues as well. I'm pretty sure that's why accelerator boards normally start at low speed then after a delay switch to high speed - the intention being to run the POST at a speed it was written for.

So could the clone board be in this scenario; the POST designed for 8088/4.77MHz operation only. The V20 runs the POST faster than it was written for resulting in the 'base 16K' (or whatever) test failing (with no indication). At that point though, assuming that the RAM refresh is still occurring, you do a quick power off/on. Due to filter capacitors on the motherboard, the switch off time might not be enough for power to be lost to the RAM, but enough for the power supply to drop/raise the Power Good line (restarting the motherboard). This time the 'base 16K' test passes.

It's just a theory. Debunking welcome.

But that doesn't explain your V20 problem on a 5160 board because I'm having no problems at all with a V20 on both the 256K (11/08/82 BIOS release) and 640K (05/09/86 BIOS release) versions of my 5160 boards.
 
Last edited:
SixPackPlus/V20/5160 incompatibility

SixPackPlus/V20/5160 incompatibility

modem7 said:
This happened to me about 6 months ago in an IBM 5160. I put the 8088 back in - system booted. Put the V20 in - system wouldn't start. Discovered that I could use the V20 only if the AST SixPackPlus board was removed.
Click to expand...

bugman2112 said:
I wonder specifically in the above case what was causing the conflict since the AST is a multifunctional card.
Click to expand...

So my earlier post where I write that I have no problems with the V20 in 5160 motherboards wasn't accurate. To be accurate, I don't see POST problems when there are no expansion cards fitted. I certainly haven't run tests to see if all BIOS calls are behaving as they should.

To the SixPackPlus/V20 issue I had seen. Six months ago, I had added the V20 into the 5160 temporarily during research into whether a driver might have have been using non-8088 instructions. I've just revisited that 5160 to see why the V20 and SixPackPlus card didn't want to work together.

Symptoms:

On first power on of 5160 - everything good.
Turn off 5160 for about 10 seconds. Power on 5160. Appears dead.
Try again: Turn off 5160 for about 10 seconds. Power on 5160. Appears dead.
Remove SixPackPlus.
Power on 5160 - everything good.
Reinsert SixPackPlus.
Power on 5160. Appears dead.
Remove SixPackPlus.
Power on 5160 - everything good.
Reinsert SixPackPlus.
Power on 5160. Appears dead.

WAIT 10 MINUTES

Power on 5160 - everything good.
Turn off 5160 for about 10 seconds. Power on 5160. Appears dead.


The only circumstance that I could get the SixPackPlus/V20/5160 combination working was when the SixPackPlus had been disconnected from power for a good 5 minutes (why doesn't 1 minute, or even 3 minutes work?).

Removed RTC chip and large WD chip from SixPackPlus. No change to symptoms.
Set switches on SixPackPlus to indicate no RAM fitted. No change to symptoms.

The 5160 I have the SixPackPlus in has a 256K motherboard that has the first (11/08/82) BIOS.
Moved the V20 and SixPackPlus into a 640K 5160 motherboard that has the last (05/09/86) BIOS - symptoms not present.

Moved the V20 and SixPackPlus back into the 256K motherboard - symptoms present (as expected).
On the 256K motherboard, upgraded the 11/08/82 BIOS to the 05/09/86 one - SYMPTOMS GONE.

So I guess I'll put this down to one of those incompatibilities that one encounters every so often: SixPackPlus (could be my particular revision) with V20 with 11/08/82 BIOS on a 256K 5160 motherboard.
 
It sounds like a reset problem. What happens if you take the dead combination, with power on, and manually reset the system by grounding the bus RESET pin (use a probe if you don't have the handy IBM reset button card)?

If it works, it probably is simple to fix.

Edit: I thought I'd add a bit of history to this. The V20 was the center of a 1989 lawsuit by Intel against NEC (Google "10 U.S.P.Q.2d 1177"). The RESET microcode was one of the main points brought out by Intel that copying had occurred, but production runs of the V20 do not have the code that Intel maintained was copied. The lawsuit was resolved in NEC's favor and is considered by IP specialists to be a key precedent when considering the issue of "expression" vs. "infringement".
 
Last edited:
Reset

Reset

After I get the system to the state where I know that turning off/on the power supply (with 10 seconds between off and on) will not work, if I then reset by momentarily grounding the POWER GOOD line, the system boots. This happened every time.

I also tried what bugman2112 did; turning the power switch off/on quickly (has to be done with just the right timing). That also worked, but I expect that that action is also effectively a reset (POWER GOOD line dips while power still present).

So an experiment for bugman2112 would be to power on his system, wait say 30 seconds (to be sure), then momentarily ground the POWER GOOD line, and see if his system then boots. But all that prove I guess is that the two actions above result in a system reset.
WARNING TO ALL:
1. Grounding the POWER GOOD line while the power supply is driving that line high might cause damage to the power supply (specifically the POWER GOOD line driver). It will depend on the circuitry used in the power supply, and on how long the grounding is done for.
2. Turning the power supply off/on quickly may stress the power supply.

Because I'm not after a fix for my scenario (V20 replace of 8088 was a temporary thing), I will now return the 8088 to the motherboard and put the 5160 back in the cupboard.
 
I'd like to try some these force resets as suggested. But I want to make sure I do it safely. I think I'm reading two different suggestions. One is to short the power good line to ground on the motherboard connector. The other is to short the reset line on one of expansion connectors. Am I correct that this is pin 28?
 
Yes, but before you do that, Modem7 may have put his finger on the problem. See if grounding the "Power good" pin on the power supply connector also does it for you.

If so, try adding a moderate-sized capacitor (say, 100 uF at 10v and observe polarity) between the Power Good line and ground at the connector and see if things improve.

The power good signal is pretty much a simple RC time-delay circuit. If it comes up too quickly because of an aged capacitor, this could be the result. Many "clone" motherboards don't use the PG line, preferring to do their own time delay on the motherboard.
 
Shorting POWER GOOD to ground resets the computer and it boots nicely. I'll dig up a capacitor and try that as you suggested here soon.
 
Couldn't get a capacitor to make the situation better. Would it be an acceptable fix to mount a momemtary pushbutton switch to short power good to reset the computer?
 
I wonder if I'm applying the capacitor correctly. Is your intention to delay the "power good" from reaching the motherboard? If so, wouldn't I want to add the capacitor in-line on the power good signal? If I put the capacitor across the "power good" and ground without interupting the other path to the motherbaord, the MB is still going to get the 5 volts of power good signal immediately, right?
 
No, it was a shot in the dark. Basically, you need to delay the onset of the power-good signal. Since IBM didn't use the same supplier for all of its power supplies, there is, to the best of my knowledge, no PSU schematic available for the 5160.

Power good can be derived in several ways, but basically it's a time-delay working off the +5. Some designs use a simple RC circuit; others use a comparator.

The permanent fix would be to dig into your PSU and find the components associated with the power-good delay and adjust them accordingly.
 
Sorry to keep this thread going, but I wanted to make sure I understood if there were any risks to the motherboard or power supply. I understand that the best possible fix is from within the power supply and I will check that out. But if I cannot figure out how to adjust power good (dang, should have finished that EE degree!), what would be an acceptable work around (if any). An external momentary switch that either shorts the +5 power good to ground? Or a momenatry switch that just interrupts power good? That would be the easiest. I suppose I could also use a junk 8 bit card to contruct a bus reset like discussed above as well. Just curious. Switching the power supply on and off twice probably isn't horrible, but those switches have some kind of finite life I would think.
 
If you wanted the simplest approach, just remove the PSU PG lead from the connector and supply a momentary contact pushbutton, one end of which is ground; the other end should be tied to the motherboard PG pin on the connector and a 1K resistor. The far end of the resistor is tied to +5. If you'd like, you can add a capacitor across the pushbutton for debouncing. 1 uF should be fine.

Your very own "push to reset" button and a mod that's completely reversible.
 
bugman2112 said:
Sorry to keep this thread going, but I wanted to make sure I understood if there were any risks to the motherboard or power supply.
Click to expand...
bugman2112 said:
An external momentary switch that either shorts the +5 power good to ground?
Click to expand...
The reset switch solution that Chuck provides (which includes disconnecting the POWER GOOD line from the PSU connector) means that the risk of damage to the PSU that I warned of earlier, is gone.

bugman2112 said:
I suppose I could also use a junk 8 bit card to construct a bus reset like discussed above as well. Just curious.
Click to expand...
For that, the RESET DRV pin (pin B2) would need to be momentarily taken HIGH. This approach has a couple of problems:
1. Depending on the motherboard circuitry, there could be a problem. For example, the 5160 motherboard will have the RESET DRV line driven low by the output of a 74LS04 (part of U71). I don't know how well a 74LS04 tolerates having its LOW output momentarily connected to +5V. Maybe short connection times are fine.
2. On a 5160 motherboard, momentarily taking RESET DRV high will only partially reset the motherboard (some chips don't get reset).

bugman2112 said:
Switching the power supply on and off twice probably isn't horrible, but those switches have some kind of finite life I would think.
Click to expand...
Unless the PSU engineers specifically designed the PSU to tolerate someone quickly powering the PSU off/on, then there is a risk of PSU damage. Different model PSUs will tolerate better than others. Some may it tolerate completely, even though not designed for such operation.
 
You must log in or register to reply here.
Back
Top