Help with identify jumping video output on VBXE/U1MB (800XL)

[chevymad]

Your tester is probably correct. You're withing 10% error. Some cap specs are as much as 50% depending on what you buy. 

[TGB1718]

3 hours ago, Beeblebrox said:

Which ground point are we talking btw, (eg related to what aspect?)

I would try to pick one of the larger ground planes on the motherboard as opposed to say a 0V line on an IC or resistor

[Beeblebrox]

1 hour ago, TGB1718 said:

I would try to pick one of the larger ground planes on the motherboard as opposed to say a 0V line on an IC or resistor

ah see what you mean. So all ground points used are large ground planes already.

[+DrVenkman]

8 hours ago, Beeblebrox said:

Case in point this brand new 2200uf rated cap comes up as 2356uf. I tried it in the top two zip slots also btw. Same reading. I know this isn't the most accurrate gadget but that is wildly out.

That's only 7% out of spec compared to the nominal 2200uF. Could be inaccuracy in the measurement, the capacitor, or most likely a little bit of each. But even assuming the meter is dead accurate, 7% variance in a power cap is *not* a big deal at all. 

[chevymad]

When I recapped a vintage reciever some of the caps listed in the service manual showed a 50% variance allowable in the specification table. I bought +/- 10% caps for the replacements. 

[Beeblebrox]

Thanks guys. New to the device and assumed it would be more accurate. Understand now how these things work.

One thing though - how am I supposed to know if a cap is failing if I can never tell using this gadget? So hypothetically if I pop in a 470uf cap and it reads between 400 and 550uf does that mean it is working? 

If I get no reading or conversely a stupidly high reading, I assume it is dead or failing respectively? 

 

 

Edited December 3, 2022 by Beeblebrox

[TZJB]

5 hours ago, Beeblebrox said:

Thanks guys. New to the device and assumed it would be more accurate. Understand now how these things work.

One thing though - how am I supposed to know if a cap is failing if I can never tell using this gadget? So hypothetically if I pop in a 470uf cap and it reads between 400 and 550uf does that mean it is working? 

If I get no reading or conversely a stupidly high reading, I assume it is dead or failing respectively? 

 

 

Technically there can be a 20% tolerance but it depends on the manufacturer's specs. If you are seeing a higher capacitance value then all the better. Can you try some other value capacitors? (Probably when your eBay order arrives)? As you have had the tester a while can I maybe suggest a firmware update for your tester? I found the original design article https://www.mikrocontroller.net/articles/AVR_Transistortester#Introduction_(English) and this useful Youtube video for an insight of an upgrade:- https://www.youtube.com/watch?v=P5zWMnvjfM4. I have only just found this so haven't done this myself yet. You do need a programmer for the ATMEGA chip though! Regarding the ground plane, one of the 130XE S-Video upgrades suggest wiring U20 CD4050BE Pin 9 to a ground plane as there is little grounding around this chip.

 

[+DrVenkman]

6 hours ago, Beeblebrox said:

Thanks guys. New to the device and assumed it would be more accurate. Understand now how these things work.

One thing though - how am I supposed to know if a cap is failing if I can never tell using this gadget? So hypothetically if I pop in a 470uf cap and it reads between 400 and 550uf does that mean it is working? 

 

 

 

First off, it’s possible the device is dead on accurate. It’s probably not, but if it’s accurate within a few percent, that’s fine. Discrete components in vintage devices weren’t manufactured THAT precisely to begin with. +/- 10% is normal manufacturing variance for a lot of consumer grade stuff of the era, and there’s a lot of tolerances throughout the system such that unless a very large percent of those components are all out of spec in the same way, everything will still work. 

 

Anyway, caps typically fail in two ways - they can short to ground, in which case you’ll get effectively no resistance between the two legs; or - if it’s an age-related degradation of the elctrolye - the capacitance goes up … a LOT. This is often (but not always) associated with swollen, bulging or leaking of the electrolyte. But if the cap is still within 10 - 15% of the nominal value, it’s fine. It’s also possible for a cap to fail open (no continuity through it at all) but this is pretty rare for an electrolytic I think.

 

 

[TGB1718]

51 minutes ago, TZJB said:

If I get no reading or conversely a stupidly high reading, I assume it is dead or failing respectively? 

Generally capacitors (non-electrolytic) go either open circuit or short circuit, electrolytics are a little

harder as under test they can appear ok, but in circuit and under load they can be "failing" although

if one is on the way out, you will probably see it's out of tolerance.

 

EDIT:- Posted same time as you @DrVenkman 

Edited December 3, 2022 by TGB1718

[TGB1718]

3 minutes ago, DrVenkman said:

but this is pretty rare for an electrolytic I think.

Yes, they usually go short and make a nice big bang as they go to meet their maker 

[TGB1718]

3 minutes ago, DrVenkman said:

but this is pretty rare for an electrolytic I think.

Yes, they usually go short and make a nice big bang as they go to meet their maker 

[TZJB]

1 hour ago, TGB1718 said:

Yes, they usually go short and make a nice big bang as they go to meet their maker 

Well that was disappointing then. R69 in my 1050 Archiver II drive siliently went dead short circuit and just killed the 5V supply 7805 regulator. All fixed now with new components!

[TGB1718]

On 12/2/2022 at 9:46 PM, Beeblebrox said:

ah see what you mean. So all ground points used are large ground planes already.

I found this, although it's talking about the de-coupling capacitors, the same rule applies when connecting

to the ground plane.

 

 

In the figure on the left (as shown above), the connection to both the power pin and the ground is made as short as possible.

It is the most effective arrangement. In the figure on the right (as shown above), the PCB trace may cause interference issues by forming a loop.

This arrangement is less effective because of the excess inductance and resistance of the PCB trace.

[woj]

@Beeblebrox any luck nailing down the audio hum issue? Asking, because my VBXE installation also has it, and it changes depending on the screen contents. I use two different power supplies, but both are switching kinds I think. No original brick in my possession to try it. I am suspecting also it could be due to something stupid I did with grounding, that I will verify shortly. The computer was fully recapped recently too.

 

EDIT: the ground connection issue I suspected is not causing it. However, looking at things I realised I included large ground and power planes on my custom adapter board, while the stock adapter does not have any, so next I need to try the stock adapter.

Edited December 12, 2022 by woj

[Beeblebrox]

38 minutes ago, woj said:

@Beeblebrox any luck nailing down the audio hum issue? Asking, because my VBXE installation also has it, and it changes depending on the screen contents. I use two different power supplies, but both are switching kinds I think. No original brick in my possession to try it. I am suspecting also it could be due to something stupid I did with grounding, that I will verify shortly. The computer was fully recapped recently too.

 

EDIT: the ground connection issue I suspected is not causing it. However, looking at things I realised I included large ground and power planes on my custom adapter board, while the stock adapter does not have any, so next I need to try the stock adapter.

Using a brick power supply over a modernn switching one helped a lot eliminating the audeio hum as well as the onscreen interferenace. Also I switched the csync pickup from the Din5 composite pin to take it from pin15 of the 4050 chip. Is that what you have done? (As I understand it you still have the 4050 chip underneath the UAV is that right?)

 

I think looking at the ground is definitely also worth it in general.

[woj]

No, no 4050 under the UAV, UAV sits directly in the 4050 socket. The stock VBXE adapter did not help either. Generally, there is interference between audio and video, for example, the booting burping sound clearly amplifies the slight jail bar effect and the video contents (for example sharp black and white screen, like the U1MB bios screen) amplifies the audio hum.

 

I have very little options for power supply, the only other one I have is a modified / stripped old PC ATX one, not even sure what kind these are, but will try it next.

 

Also, one of the ground connection required on one of the VBXE headers I took from a place that is a bit doubtful, now that I think of it...

[_The Doctor__]

personally I'd put the 4050 back in, pull from there as suggested, then I'd look for grounding and cable shielding issues along with making sure the power supply is fully filtered and clean

[woj]

Yes, I think I need to try that, connecting it to my ATX power supply seems to have helped a little bit, but then this is very subjective, there is no way to measure it in absolute terms.

[woj]

I did try (and fixed permanently) 4050 underneath the UAV and hooking up CSYNC directly to 4050, no profound difference. But in general it is manageable, with normal audio levels the hum is not really audiable. 

 

But in any case, I am fresh to this new age of Atari computing, I left them back in around 1995 when my first better PC came, it is astonishing to me how many different quirks these machines have, and also that it was acceptable back in the day. I guess the fact that they were relatively cheap helped a lot, and also, had they been designed better they would have needed to cost much much more too. I also wonder, zero experience here, are the new remakes, like the NUC or the 1088XEL, are they free of such things, or is the same story all over again?  

[Beeblebrox]

AFAIK back in the day personal computing was in its infancy (relatively speaking) and a lot of Atari's customers were just plugging these mass produced 8-bits into TVs via rf. So quality control for these I'd imagine wasn't high on Atari's priority. The XEs were certainly not as well built as the XLs it seems. 

[woj]

So I made a discovery today about the audio hum, when I have the standard DIN-5 video cable also hooked up to the Atari and the screen on the other side, in parallel with the DIN-13 Coolnovelties cable attached to the same screen, then the hum on the DIN-13 input on the TV is considerably reduced. Thus, I think, it has something to do with the shielding, as the audio cable I have for the DIN-5 cable is properly shielded. I am yet to check if the CN cable is shielded at all, but by the looks of it / touch, it is not, so perhaps something to be gained on this front... would hate it though if it means that I have to rewire the whole CN cable, not looking forward to it at all.