Maker Pro
Maker Pro

Looking for a replacement of NE38B Neon Indicator Bulb

Hi All,

This is my first post here (sorry for the long one) and I am rather a newbie to electronics, so please be patient with me.

As one of my first projects, I picked a quite difficult one. I purchased very cheaply old oscilloscope Aaron BS-635 (also known as the later rebranded Hung Chang OS-635). I simply can't afford to spend $400+ on a new or used piece. The oscilloscope had a blown fuse and I was told that prior to that the CRT lit up for a short while and then the fuse blew.

I decided to give it a go. After all, "if people made it, people can fix it" :)

I replaced ALL the electrolytic caps throughout the oscilloscope (what the heck - they were only 30 or so :) ), as most of them were old and this is the first thing which normally gives up the ghost, I was told.

I managed to connect everything correctly (or at least as it was before I took it apart) with the help of the photos I was taking before each step of the way. I also labelled all the cables with the source and destination connectors numbers for a future reference.

While working on the Power Supply board, I also discovered an internal fuse, which was blown. Of course, I replaced it as well.

When I powered up the oscilloscope for the first time after assembling it, the CRT came alive!

But there was a lot of sparks going between the 4.5kV CRT cable mount (solder) on a high voltage PCB and the internal high voltage enclosure (grounded). There is a distance of about 8-10mm between the PCB and the enclosure.

Since there were already some parts missing when I bought it (e.g. the high voltage enclosure lid, as well as some plastic shields here and there - I found one loose inside not in its place and I can't find where it belongs), I thought of simply shielding/insulating between the PCB and enclosure.

This helped - at least to eliminate sparks going there. But they started to form between the legs of the high voltage ceramic (metal film?) caps! The legs were originally insulated in most parts by some sleeves, but there were tiny exposed lengths just at the PCB connections. Using insulating liquid ("liquid insulating tape"), I made sure to cover these tiny bits. This helped too.

But... then the sparks started to fly between the CRT 4.5kV cable itself (through its insulation!!!) and.. the tops of the caps!

That was the moment when the internal (high voltage) fuse decided it had enough.

I took out the PCB again, replaced the fuse and started to examine the PCB under the magnifying glass.

The first thing I noticed, the CRT cable insulation had several micro cracks (age), so I covered all of it in several layers of "liquid tape". Then I noticed that one of three Neon Indicator Bulbs had one leg missing (poor thing).

Here starts my problem. The original neon bulbs (N1, N2 and N3 on the schematic) are type NE38B. These are small "pigtail" tubes, about 5mm long (picture below). They are connected in series. Unfortunately, I can't find anywhere not only any replacement but even any data about them. Apparently, they don't exist. I also don't have any clue at what voltage they are rated, as the schematic doesn't mention any voltage on the rail they are on. The cap next to them is rated at 500v, so my guess is they should be similar. The highest rating I found the new ones available for is 250v. I don't know virtually anything about the neon bulbs (just read the basics today). The fact that they are connected in series is probably very important - does it mean that they could individually be a lower rating, so they "make it up in the gang"? I think that in series they work like resistors, so this would be my guess.

Could this be a reason for the sparking (too high voltage because of the faulty neon bulb)?

I haven't tested the actual voltage because I have no meter able to test 4.5kV...

The black top on the neon bulb (photo below) is a little bit of the "liquid tape" (by accident) but is irrelevant to the problem.

If anybody could help me with any comments/ideas, I would be extremally grateful!

NE38B.jpg AaronSchematic.png
 
Welcome to EP!
Looks like the three neons together are acting as a voltage reference (about 3 x 80V = 240V) for voltage limiting/stabilising purposes. If you can't get exact replacements you could try common mains indicator neons.
 
veavd
Welcome to EP!
Looks like the three neons together are acting as a voltage reference (about 3 x 80V = 240V) for voltage limiting/stabilising purposes. If you can't get exact replacements you could try common mains indicator neons.
Thank you for your help. It confirms my guess that neons (like resistors) in series add their value, but to what voltage? I'm not sure if 240V is enough - this is a high voltage circuit and surrounding caps are rated for 500V to 2kV. Two series resistors in parallel to the neons are 22MΩ each. This is heavy stuff and without any hint to the rail voltage, I can only guess the kind of magnitude there. I could perhaps try to analyse the circuit and attempt to calculate voltage at that point, but with my VERY limited knowledge and experience, I would not trust myself. I might resort to redrawing the circuit in the software (at least from the known voltage points) and spice it at this rail. Sound like fun :).

Now I understand that the sparking could be related to the neons indeed. If they are in a series and one doesn't work (open circuit) none of them works. They theoretically don't exist in the circuit. So, Heavens know what voltage was shooting through that CRT wire and I only hope I didn't fry the tube itself. On the other hand, it looks like this problem was present for some time (certainly since I bought the oscilloscope) and CRT survived (it lit up several times) during my tests (until the last one when something else might have died), so there still might be a hope... :)

Any more comments from anybody would be much appreciated. With your help, I may make my newbie self proud for fixing this oscilloscope (not to mention its usefulness to me) :)
 
Check your 22MΩ resistors. There will be a feedback circuit somewhere to set the voltage.

If the EHT circuit is producing AC rather than DC then nasties will occur. Check rectifiers and reservoir capacitor.
 
Check your 22MΩ resistors. There will be a feedback circuit somewhere to set the voltage.

If the EHT circuit is producing AC rather than DC then nasties will occur. Check rectifiers and reservoir capacitor.
Thanks duke37 for your comment. Looks like I will have to Google your answer because right now I need to be spoken to VERY slowly :) I am a newbie to this stuff :) I am not sure what do you mean by "Check your 22MΩ resistors". Should I check if they are OK? With the tester? I will do it. Rectifiers I understand but again, should I just check them with the tester in the diode mode if they behave correctly? I will do that too :) From what I understand, a reservoir cap in the rectifier system is usually an electrolytic cap (at least in analogue circuits). Am I correct? I replaced all of the electro caps with the brand new ones throughout. But I will check these anyway (there are several rectifier systems there).

I am not sure how to recognise a "feedback circuit" but I will Google this.

All-in-all, thank you for your help. Please keep them coming! :)
 
High value resistors can often go up in value, if these are in the feedback, then more voltage than intended will be generated.
If the neons are the voltage reference and it is much too high because they are open circuit, then the output voltage will be too high.
I did not spot the rectifiers but you may not be able to test these with a normal tester since they probably contain several diodes in series and need several volts to turn on.
The EHT reservoir capacitor is not likely to be an electrolytic, it is more likely to be a high voltage ceramic. What is the frequency of the EHT generator, is it from the mains or from a high frequency invertor?

A feedback circuit is used to adjust the output to a specific value even though the load and input may vary. An op-amp is unlikely to be used in a high voltage circuit. What is done is to compare the output to a reference and adjust so they match.
 
Top