Phillips CM8833 Troubles

for bbc micro/electron hardware, peripherals & programming issues (NOT emulators!)
Post Reply
Kazzie
Posts: 282
Joined: Sun Oct 15, 2017 7:10 pm
Location: North Wales
Contact:

Phillips CM8833 Troubles

Post by Kazzie » Sun Sep 16, 2018 2:48 pm

My monitor just gave up on me.

It was connected to my A420/1, displaying a static screen, when the picture spontaneously disappeared, the power LED extinguished, and the normally very-very high-pitched whine plummeted significantly. By the time I got my phones spectrum analyzer on it, it was ringing around 2.5kHz, and is dropping steadily over time (whether remaining switched on or left off for a while). The electron gun doesn't appear to be firing anything at the screen. No funny smells have escaped, and the box is warm, but not unusually so.

I'm not going to have the time to open her up before next weekend, but does anyone have any pointers on where I need to look inside? (While I've discharged tubes and worked inside CRTs before, it's always been on the surrounding gubbins, as opposed to the HV stuff itself.
Last edited by Kazzie on Sun Sep 16, 2018 2:49 pm, edited 2 times in total.
BBC Model B 32k issue 7, Sidewise ROM board with 16K RAM
Archimedes 420/1 upgraded to 4MB RAM (mid- restoration)
Acorn System 1 home-made replica

User avatar
AndyF
Posts: 1090
Joined: Sat Feb 23, 2008 10:16 pm
Location: Derby
Contact:

Re: Phillips CM8833 Troubles

Post by AndyF » Sun Sep 16, 2018 3:49 pm

Without specific knowledge of this monitor although I've seen them in years past (and they are a decent thing usually with Stereo out in some cases, at least the ones I saw commanded relatively high prices in good nick) , the biggest clue I saw in your post was the fact the power light / neon / led (whatever it is) went out.

To my initial thoughts of mind, this points towards a problem with the initial power supply circuitry hopefully, possibly even the LV (12V? maybe) output for the audio/power led etc/ , I'm mainly thinking here of a open circuit rectifier or cap perhaps. I guess all I'm saying is with the description I'd expect to find the 'concern' near the mains input area not further down or near the HV area**

** HV as in "more than mains aka tripler area
Last edited by AndyF on Sun Sep 16, 2018 3:50 pm, edited 1 time in total.
Andy

* NEW * The Jetset Willy and Manic Miner community :)

Adventure games ported across to the BBC (in progress) as soon as I can find some time!

Kazzie
Posts: 282
Joined: Sun Oct 15, 2017 7:10 pm
Location: North Wales
Contact:

Re: Phillips CM8833 Troubles

Post by Kazzie » Sun Sep 16, 2018 4:10 pm

AndyF wrote:
Sun Sep 16, 2018 3:49 pm
Without specific knowledge of this monitor although I've seen them in years past (and they are a decent thing usually with Stereo out in some cases, at least the ones I saw commanded relatively high prices in good nick) , the biggest clue I saw in your post was the fact the power light / neon / led (whatever it is) went out.

To my initial thoughts of mind, this points towards a problem with the initial power supply circuitry hopefully, possibly even the LV (12V? maybe) output for the audio/power led etc/ , I'm mainly thinking here of a open circuit rectifier or cap perhaps. I guess all I'm saying is with the description I'd expect to find the 'concern' near the mains input area not further down or near the HV area**

** HV as in "more than mains aka tripler area
That matches my suspicions, having had a quick general read on the interweb about CRT issues. (This one is a model with Stereo output, and a hand-me-down from my old high school.)

There's a service manual available online, which contains a circuit diagram, component list, and instructions for calibrations after repair (but the technician is expected to be able to repair it from the diagram).

I'll gladly accept any further guidance or pearls of wisdom from those who opened up CRTs back in the days when looking at an LCD display meant you were using a calculator.
BBC Model B 32k issue 7, Sidewise ROM board with 16K RAM
Archimedes 420/1 upgraded to 4MB RAM (mid- restoration)
Acorn System 1 home-made replica

User avatar
AndyF
Posts: 1090
Joined: Sat Feb 23, 2008 10:16 pm
Location: Derby
Contact:

Re: Phillips CM8833 Troubles

Post by AndyF » Sun Sep 16, 2018 4:45 pm

Someone else will be able to advise on that point I expect. I looked at the manual and thankfully the power supply diagram is separate from the rest, the bottom part of Page-16 , if it was me that is where I would start, being aware that mains voltage or higher is exposed , usual safety precautions apply here, main one being if in doubt = don't! , I know you probably know that but I have to type it for my peace of mind sorry. :oops:

:)

One thing that might help others to help you , what do you have in terms of testing kit ? I'm assuming a multimeter, do you have a 'scope or a logic probe / anything else ?
Andy

* NEW * The Jetset Willy and Manic Miner community :)

Adventure games ported across to the BBC (in progress) as soon as I can find some time!

User avatar
1024MAK
Posts: 7994
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Phillips CM8833 Troubles

Post by 1024MAK » Sun Sep 16, 2018 5:05 pm

As with anything else, after a good visual inspection, the next step is to work out (by testing) what power supply rails are okay and which ones are missing or low. No need to worry about testing the EHT though.

If the tube heater is not glowing, that normally means it has lost it’s supply.

Mark

Kazzie
Posts: 282
Joined: Sun Oct 15, 2017 7:10 pm
Location: North Wales
Contact:

Re: Phillips CM8833 Troubles

Post by Kazzie » Sun Sep 16, 2018 6:34 pm

AndyF wrote:
Sun Sep 16, 2018 4:45 pm
Someone else will be able to advise on that point I expect. I looked at the manual and thankfully the power supply diagram is separate from the rest, the bottom part of Page-16 , if it was me that is where I would start, being aware that mains voltage or higher is exposed , usual safety precautions apply here, main one being if in doubt = don't! , I know you probably know that but I have to type it for my peace of mind sorry. :oops:

:)

One thing that might help others to help you , what do you have in terms of testing kit ? I'm assuming a multimeter, do you have a 'scope or a logic probe / anything else ?
Multimeter, check. My el-cheapo oscilloscope died a few weeks ago, but I can cart the monitor in to work and use a 'scope there (as long as I don't tell the H&S people what I'm repairing).
BBC Model B 32k issue 7, Sidewise ROM board with 16K RAM
Archimedes 420/1 upgraded to 4MB RAM (mid- restoration)
Acorn System 1 home-made replica

Kazzie
Posts: 282
Joined: Sun Oct 15, 2017 7:10 pm
Location: North Wales
Contact:

Re: Phillips CM8833 Troubles

Post by Kazzie » Sun Sep 16, 2018 6:50 pm

1024MAK wrote:
Sun Sep 16, 2018 5:05 pm
As with anything else, after a good visual inspection, the next step is to work out (by testing) what power supply rails are okay and which ones are missing or low. No need to worry about testing the EHT though.

If the tube heater is not glowing, that normally means it has lost it’s supply.

Mark
I've just plugged the monitor back in for two minutes, and while it's still buzzing at me, I can't detect any heat convecting out of it.

It seems that this failure may not be as fatal as I initially feared. Thank you for your pointers and reassurance.
BBC Model B 32k issue 7, Sidewise ROM board with 16K RAM
Archimedes 420/1 upgraded to 4MB RAM (mid- restoration)
Acorn System 1 home-made replica

Kazzie
Posts: 282
Joined: Sun Oct 15, 2017 7:10 pm
Location: North Wales
Contact:

Re: Phillips CM8833 Troubles

Post by Kazzie » Fri Sep 28, 2018 8:46 am

I've finally found the time to take screwdriver and multimeter to the monitor. Here are my initial findings:

The high-pitched whine is coming from the power supply board, specifically the large T101 power transformer. (I stuck a funnel on a plastic tube to use as a stethoscope!) All the components on board are in a visually good condition (with the possible exception of one capacitor which may have a slight bulge, but there's no visible leakage at all).

At first I thought the connector M2 that provides 125V, 25.5V and 16.1V to the main board was giving no output at all, and I proceeded to probe the individual stages of the board, but on double-checking last thing yesterday it gives reasonable voltages when run disconnected from the main board, with no load (e.g 140V, 28V, 18V or somewhat). I may have picked a ground point covered in flux the first time, or it may be behaving differently when disconnected. Though if it is behaving different when connected, I didn't notice any difference in the pitch of the whine, which did vary as I probed with my multimeter. (Likewise, some of my cheaper wall-wart power supplies change pitch when their load is added/removed.)


There's clearly something odd in my diagnosis, as the LED is powered directly from the 16V supply of the main board, so if the power supply was indeed giving 16V to the board the LED should light up. Unless it got fried when the monitor died? I'll look again when I have a clear table and the kids are in bed.
BBC Model B 32k issue 7, Sidewise ROM board with 16K RAM
Archimedes 420/1 upgraded to 4MB RAM (mid- restoration)
Acorn System 1 home-made replica

User avatar
1024MAK
Posts: 7994
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Phillips CM8833 Troubles

Post by 1024MAK » Fri Sep 28, 2018 11:14 am

The ‘whine’ that a SMPSU makes can change depending on load. It depends on what type of SMPSU it is. Some normally run at a nominal ‘fixed’ frequency and the on time of the switching transistor is varied (mark-space ratio changes). While some vary the frequency of operation (within a limited range).

However, most will sound different if either, they have gone into protection mode due to detecting an overload or short circuit on one of their outputs, or the normal minimum load is not connected.

Mark
Last edited by 1024MAK on Fri Sep 28, 2018 1:40 pm, edited 1 time in total.

Kazzie
Posts: 282
Joined: Sun Oct 15, 2017 7:10 pm
Location: North Wales
Contact:

Re: Phillips CM8833 Troubles

Post by Kazzie » Fri Sep 28, 2018 12:25 pm

1024MAK wrote:
Fri Sep 28, 2018 11:14 am
However, most will sound different if either, they have gone into protection mode due to detecting an overload or short circuit on one of their outputs, or the normal minimum load is not connected.
Which is what I think is happening here. (This particular sound first appeared when the monitor failed, so that would indicate it's in protection mode.)
BBC Model B 32k issue 7, Sidewise ROM board with 16K RAM
Archimedes 420/1 upgraded to 4MB RAM (mid- restoration)
Acorn System 1 home-made replica

User avatar
DutchAcorn
Posts: 1899
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands
Contact:

Re: Phillips CM8833 Troubles

Post by DutchAcorn » Fri Sep 28, 2018 12:31 pm

Hi high pitched whine and no power led are also the symptoms of a dead lopt; probably the most common cause of failure for these monitors.
Paul

Kazzie
Posts: 282
Joined: Sun Oct 15, 2017 7:10 pm
Location: North Wales
Contact:

Re: Phillips CM8833 Troubles

Post by Kazzie » Sun Sep 30, 2018 9:25 am

Another small step: measuring resistance across the power supply board's outputs, when connected to the main board and isolated, I found little change in the 16V and 26V lines (Kohm or Mohm ranges), but the 125V dropped to 56 ohms(!) when connected to the main board.

Working from the circuit diagram, the 125V supply connects to two points: R469 next to the line output transformer, and through connector M6 to the electron gun's board. Having disconnected M6 the resistance on the 125V output of the power supply was still 56 ohms, which indicates the fault is on the main board near the lopt (or the lpt itself).
cm8833-125v.png
R469 is a beefy 12ohm, 5W resistor, and measures fine at 12ohm. The remaining ~44ohm shows up across C469 (and C470). Snooping around nearby, the resistance around R466 (68ohm, underlined in blue), is 5.8ohm. That leaves me looking suspiciously at the BU508A transistor next to it. Reading across it's pins I find CB=50ohm, CE=44.6ohm, BE=5.8ohm.

Conveniently, the BU508A (TS467) is right next to the edge of the board, so should be straight-forward to remove from the board for testing. But not right now: it's bath time for the kids...
Last edited by Kazzie on Sun Sep 30, 2018 9:26 am, edited 1 time in total.
BBC Model B 32k issue 7, Sidewise ROM board with 16K RAM
Archimedes 420/1 upgraded to 4MB RAM (mid- restoration)
Acorn System 1 home-made replica

User avatar
1024MAK
Posts: 7994
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Phillips CM8833 Troubles

Post by 1024MAK » Sun Sep 30, 2018 9:40 am

It may be that the line output transistor BU508A (TS467) failed on it’s own. Or it may have overheated due to a problem elsewhere. If the LOPT provides auxiliary supplies, do a resistance check on these. Also test diodes D467 and D468.

Fit a new one and see how long it lasts while you check it’s base drive waveform. Test the voltages of any LOPT provided auxiliary supplies.

Mark
Last edited by 1024MAK on Sun Sep 30, 2018 9:42 am, edited 2 times in total.

Kazzie
Posts: 282
Joined: Sun Oct 15, 2017 7:10 pm
Location: North Wales
Contact:

Re: Phillips CM8833 Troubles

Post by Kazzie » Sun Sep 30, 2018 5:51 pm

1024MAK wrote:
Sun Sep 30, 2018 9:40 am
It may be that the line output transistor BU508A (TS467) failed on it’s own. Or it may have overheated due to a problem elsewhere. If the LOPT provides auxiliary supplies, do a resistance check on these. Also test diodes D467 and D468.

Fit a new one and see how long it lasts while you check it’s base drive waveform. Test the voltages of any LOPT provided auxiliary supplies.

Mark
Thanks. D467 and D468 both give a sensible 0.6V drop on a diode test.

Having peered a little closer throughthe maze of wires, I've found that my LOPT (HR7506) isn't the exact same as that specified in the circuit diagram I have (HR6489). It seems that the two versions of the CM8833 are similar enough that I hadn't noticed the difference until now! Here's a manufacturer's schematic of my LOPT:

Image

(The key differences with the HR6489 are a different physical footprint and lack of focus unit in the latter.)

There are no auxiliary power supplies to my eye. The 30V PTP pulse output goes round the RGB elements of the electron gun via two inductors, so it seems reasonable for it to offer no resistance.

I desoldered the BU508A and tested it out of circuit, and found that it was giving sensible resistances between the pins. The 5~50ohm resistances remained on the board, and there was no change to the resistance seem by the power supply. As the soldering iron was hot, I progressively removed or isolated other components, namely C467-471, C494, R465-466, R494, D465, D467 and D468. (I noted that I'd misread R465's value, it's 3R3, not 3K3, which means that a 5ohm resistance over R466 is understandable.) That effectively left the 125V supply connected (through R469) to only the LOPT, and the power supply still saw ~56ohms on the 125V line. Isolating R469 left the 125V pin floating, and the ~56ohm reading remained between the LOPT's pins and ground: pins 18 and 6 in the circuit diagram (but 9 and 10 on my actual LOPT).

With all those surrounding components removed, I did a continuity test on the LOPT's pins. I found the following groupings: 1-3, 2-5-6-9-10-11, 7. (On the circuit diagram, that's 1-5, 4-12-8-6-18, 11.) 2-5-6 shouldn't have continuity to 9-10, cause 10(18) should be at 123V according to the circuit diagram, and 5(12) is grounded! Given that all components connected to pins 9 and 10 had been disconnected, that means that there must be a short circuit in the LOPT.

It looks like replacements are available online for ~£40 or so. So now that I know exactly what is needed, all I need is to wait for payday... :roll:
BBC Model B 32k issue 7, Sidewise ROM board with 16K RAM
Archimedes 420/1 upgraded to 4MB RAM (mid- restoration)
Acorn System 1 home-made replica

User avatar
DutchAcorn
Posts: 1899
Joined: Fri Mar 21, 2014 9:56 am
Location: Maarn, Netherlands
Contact:

Re: Phillips CM8833 Troubles

Post by DutchAcorn » Mon Oct 01, 2018 6:00 am

Thanks for sharing these details. Having had similar issues I read “somewhere” that a LOPT in itself is quite robust and usually fails due to stress caused by degraded related components.

Not knowing what related components to look for I replaced the LOPT on mine which brought it back to life. But it died again within a week (with other symptoms). While you are analysing it may be a good idea to do a good check on the components that the LOPT connects to.
Paul

Kazzie
Posts: 282
Joined: Sun Oct 15, 2017 7:10 pm
Location: North Wales
Contact:

Re: Phillips CM8833 Troubles

Post by Kazzie » Mon Oct 01, 2018 2:43 pm

I'm glad you've found it useful. I've used this post as a mix of looking for advice from those with experience, a chance to organise my thoughts by writing them down, an aide memoire for what I've already done, and some possibly useful reading for someone on the internet that has a similar problem in the future.

The components I pulled out yesterday account for a lot of those connected to the LOPT. Most of the remaining ones are on the electron gun board. I'll also probably replace the BU508A anyway: it's only a pound or two, a pittance next to the cost of the LOP, and as I said it's in a really accessible location to desolder again.
BBC Model B 32k issue 7, Sidewise ROM board with 16K RAM
Archimedes 420/1 upgraded to 4MB RAM (mid- restoration)
Acorn System 1 home-made replica

Kazzie
Posts: 282
Joined: Sun Oct 15, 2017 7:10 pm
Location: North Wales
Contact:

Re: Phillips CM8833 Troubles

Post by Kazzie » Wed Oct 31, 2018 10:20 am

I finally fitted the new LOPT and line output transistor last night, to some partial success.

The tube warms up, brightening from black through gray to white, then apparently cuts out back to black, and starts the loop again. I've not left it running for more than a few seconds at a time because of this.

I'm pretty sure that all the passive components that I (partially) desoldered to isolate for testing were put back in properly. It seems I've got more testing to do.
Last edited by Kazzie on Wed Oct 31, 2018 10:21 am, edited 1 time in total.
BBC Model B 32k issue 7, Sidewise ROM board with 16K RAM
Archimedes 420/1 upgraded to 4MB RAM (mid- restoration)
Acorn System 1 home-made replica

User avatar
1024MAK
Posts: 7994
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Phillips CM8833 Troubles

Post by 1024MAK » Wed Oct 31, 2018 10:39 am

I don’t know about this chassis, but if the electron gun current gets to too high a value, a protection circuit will cause a shut down. The less complex protection circuits don’t latch, so the set will then cycle. So how bright is the white screen?

Mark
Last edited by 1024MAK on Wed Oct 31, 2018 10:40 am, edited 1 time in total.

Kazzie
Posts: 282
Joined: Sun Oct 15, 2017 7:10 pm
Location: North Wales
Contact:

Re: Phillips CM8833 Troubles

Post by Kazzie » Thu Nov 01, 2018 10:45 am

1024MAK wrote:
Wed Oct 31, 2018 10:39 am
I don’t know about this chassis, but if the electron gun current gets to too high a value, a protection circuit will cause a shut down. The less complex protection circuits don’t latch, so the set will then cycle. So how bright is the white screen?

Mark
Your instincts were right again. After a spot of (further) background reading, I tried adjusting the "SCREEN" potentiometer on the LOPT. The display now powers up and stays on: the pot was adjusted near its maximum brightness at the factory, which led to the tripping.

Linked up to my Beeb I got a blurry image of my Beeb's (composite) video output. So after adjusting the "FOCUS" pot, we're back in business.

Thank you very much for your encouragement and guidance along the way.

(The screen was going all the way to "full white", which given that it had no input connected, was far more than it should have.)
BBC Model B 32k issue 7, Sidewise ROM board with 16K RAM
Archimedes 420/1 upgraded to 4MB RAM (mid- restoration)
Acorn System 1 home-made replica

User avatar
1024MAK
Posts: 7994
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Phillips CM8833 Troubles

Post by 1024MAK » Thu Nov 01, 2018 1:40 pm

Well done =D>

It’s good that you have got it working again. Long live CRT technology :D

Mark

Post Reply