• Please review our updated Terms and Rules here

Model 1 coposite video.

P

prime

Experienced Member
Joined
Sep 20, 2009
Messages
153
Location
Coventry, UK
Hi All,

I've got my model 1 working now, and upgraded internally to 48K (using 4164 RAMS), which means that I can run the machine on a simple regulated 5V supply.
This all seems to be working properly and I have 48K of RAM available.

However the display looks a little "unsteady" some adjacent scan lines seem slightly offset from one another maybe a pixel or so. Also I seem to be getting odd pixels missing for some reason.
I'm using the composite input of an LG 17" flat screen TV that has worked correctly with most of my other retro collection, so that should be ok.

Anyone know what the problem may be?

Cheers.

Phill.
 
Is it a flat CRT or a flat panel, e.g., LCD? I don't believe composite is always a no-brainer with a flat panel monitor.

Did the other machines you used with it supply composite output or is the Model 1 the only composite source you've used with it?
 
Stone said:
Is it a flat CRT or a flat panel, e.g., LCD? I don't believe composite is always a no-brainer with a flat panel monitor.

Did the other machines you used with it supply composite output or is the Model 1 the only composite source you've used with it?
Click to expand...

Monitor is an LCD.

Have used it with other machines on Composite, Tandy CoCo, Dragon 32/64, Acorn Atom, BBC Micro, Commodore 64, Atari 800 etc.

Cheers.

Phill.
 
prime said:
Hi All,


However the display looks a little "unsteady" some adjacent scan lines seem slightly offset from one another maybe a pixel or so. Also I seem to be getting odd pixels missing for some reason.
I'm using the composite input of an LG 17" flat screen TV that has worked correctly with most of my other retro collection, so that should be ok.

Anyone know what the problem may be?

Cheers.

Phill.
Click to expand...

Phill
Do you have a CRO to look at the signal " Level " that is being fed to the Composite input of the TV.
Thanks Ray
 
Is your +5V supply coming from a switchmode adaptor?

If so that will be your problem. I tried to upgrade mine to a switchmode and I had the same problem (unsteady video, video noise etc). I went back to the linear regulator and everything worked fine again. The main reason I did this was the same as you - fitted 4164's and didn't need the other voltages (see caveat below).

The problem is that the HF harmonics in the switchmode power supply aren't filtered out properly inside the TRS80 as it wasn't designed for them. You might be able to do something by fitting a LC filter with a low cutoff frequency between the regulator and the TRS; however I didn't bother with mine as the old linear supply works fine.

Caveat - On my system (which is a clone) the tape relay was fed with the +12 supply. Without the +12v it won't work. This may or may not be the same for the real deal TRS80.
 
Right I've done some more investigating this morning....

I've replaced the Electrolytic caps that are near the video section but that didn't seem to help.

Then I noticed something, occasionally the machine will start up with just a screen full of junk and the CPU doesn't appear to be running. When it starts like this the display is rock solid.

So I pulled the CPU and started without it plugged in, again rock solid. So then I started the machine up as normal, and hit enter at the memory size prompt, to get into basic. This had the unsteady video, but when I grounded the /wait line of the CPU and stopped it, the video stopped being unsteady and was stable. Needless to say the unsteadyness in the video seems to be related to the CPU running.

This also seems to be unrelated to which Z80 is connected, as whilst bringing the machine up I had tried a Z80A as I was unsure if the CPU was OK, and this also had the video problem.

Yep switchmode adapter, but never had a problem with it on other machines, I also scoped it's output and it is a solid 5V with no interference on it that I can see.
Tape relay is +5V on the TRS80.

Cheers.

Phill.
 
prime said:
Yep switchmode adapter, but never had a problem with it on other machines, I also scoped it's output and it is a solid 5V with no interference on it that I can see.
Tape relay is +5V on the TRS80.

Cheers.

Phill.
Click to expand...

It really doesn't take much to cause problems, the noise in my system was < 10mV in the kHz range. It wasn't able to be corrected with an array of electrolytics, hence why I suggested an LC filter.

Did you remove all the bypass caps from the +5V rail around the 4164 RAM chips (which is now one of the data lines for the 4164's?) If they are still present then it will cause all sorts of unpredictable behaviour. Also note that now your old +12V rail is now a +5V rail so it will may be lacking enough bypass capacitors; again causing unpredictable behaviour.

Also how are you hooking up your new +5V supply? Did you simply tie it to the +5V rail internally or is it still being fed through the internal regulators?

EDIT: Finally are your 4164 IC's 7-bit refresh compatible? If they aren't then you'll have all sorts of weird things going on.
 
Last edited:
Hi All,

I've now also swapped out all the decoupling caps in case they had gone bad, and I've uploaded a video of the problem here : https://youtu.be/r03GOJwv_os

I've made sure there is a decoupling cap for each of the 4164s, and yep they are 128 cycle refresh chips (well according to the datasheet, and yeah it's the exact datasheet for the chips, same manufacturer serial no etc....).
Also tried with a 5V linear supply and that also has the same problem.

There seem to be two ways I can make the problem go away :
1) as previously mentioned halt the CPU.
2) By winding R20 all the way over so it's effectively a 0K resistor, but this results in the display being moved too far to the right and so about the last 10 columns are off the screen.

Cheers.

Phill.
 
Last edited:
It looks like there is a DC Shift / Drift / Noise on the Output of Q1. ?
Or you could put a !0 MF Cap u in series with the video line to isolate the dc shift ?
Ray
 
Last edited:
Hi all,

Just to let people know that I eventually found a solution after trying to re-engineer the sync circuit (on a breadboard) I noticed that the +5V rail was very noisy so connected a 1000µf cap across it. That made an immidiate improvement. So I disconnected my re-engineered sync, and went back to the components on the motherboard but then connected the cap across Z6, and another across Z5, and this seems to have completely solved the problem.

Cheers.

Phill.
 
Pleased you've got it sorted Phil. The video reminded me of the interference the screen would get, in the old days, if it was too close to the disks or printer
 
It looks like your old +5V rail to the ram sockets (now rewired as an address line) is not completely isolated from the remaining +5V rail.

I would say it's either the trace location in proximity to something 'sensitive' (i.e. now putting noise on it) or there is some passive that is still connected that shouldn't be (i.e. filter cap, resistor to somewhere etc).

Either way it looks like it's working again; well done.
 
You must log in or register to reply here.
Back
Top