Tektronix TDS210 repair - oscillator help needed

[al_m]

Hello all,

I found this topic after searching for TDS210 repairs, and thought i'd see if anyone has any advice for my next move in repairing a TDS210 i picked up with the blank display issue.

I have verified all the power rails are good, but the LCD flex connector was missing the two negative voltages on pins 1+2. The data lines also appear to be pinned high, but i can't verify this as i don't have another scope to check with! The upper 4 read 5.1VDC, or ~10VAC on my meter, which seems odd. Anyway, i went ahead and stuck in a bodge wire connecting pin 2 on the LCD connector straight to the -24V off the power connector, and pin 1 to the same rail, through a 100k resistor. Powering up after that gave me -22V on LCD pin 1, and -24V on pin 2, so within spec for the LCD according to the service manual. After plugging in the display i now get a black border around the screen! Some progress, but nothing more happens.

Is it likely then to be a fault in the main oscillator, as the first poster reported with his unit? Is there some way the circuitry is switching in the voltages for the LCD after boot, which isn't happening cos the osc is buggered?

Any tips would be much appreciated. I'm pretty limited in my tool set here, my main scope is about 120 miles away and we're in lockdown, so no go. Thinking of picking up a second unit just to be able to probe this one but am a bit skint at the mo.

[george.b]

Hello all,

I found this topic after searching for TDS210 repairs, and thought i'd see if anyone has any advice for my next move in repairing a TDS210 i picked up with the blank display issue.

I have verified all the power rails are good, but the LCD flex connector was missing the two negative voltages on pins 1+2. The data lines also appear to be pinned high, but i can't verify this as i don't have another scope to check with! The upper 4 read 5.1VDC, or ~10VAC on my meter, which seems odd. Anyway, i went ahead and stuck in a bodge wire connecting pin 2 on the LCD connector straight to the -24V off the power connector, and pin 1 to the same rail, through a 100k resistor. Powering up after that gave me -22V on LCD pin 1, and -24V on pin 2, so within spec for the LCD according to the service manual. After plugging in the display i now get a black border around the screen! Some progress, but nothing more happens.

Is it likely then to be a fault in the main oscillator, as the first poster reported with his unit? Is there some way the circuitry is switching in the voltages for the LCD after boot, which isn't happening cos the osc is buggered?

Any tips would be much appreciated. I'm pretty limited in my tool set here, my main scope is about 120 miles away and we're in lockdown, so no go. Thinking of picking up a second unit just to be able to probe this one but am a bit skint at the mo.

Could you post a picture of what you're getting on the display?

[al_m]

Could you post a picture of what you're getting on the display?

Hi George, yes, here's a picture.

There's definitely life compared to what it was like before (totally blank). Now when i turn the unit off i can see the scan line die out, and it's in a different position every time. So there is probably a clock running somewhere. But, having read more into it i realised i was setting the contrast by putting a voltage on pin 1, so i removed my bodge resistor and it made no difference. There does seem to be a signal on the probe comp lug, i get ~2.5VDC or ~6VAC when i stick my meter across the terminals, which sounds about right for a 5V 1kz square wave. Hmm.

[george.b]

Alright, so, the black border is normal. I wasn't entirely sure what you meant, I thought it was a black border inside the display area, but that's outside it, so there's nothing wrong with that.

Ideally you'd check the probe comp signal at least with a frequency meter, since you don't have a second scope handy. Maybe you could cobble one up with a microcontroller if you have access to any. The TDS200 series will output a square wave of different frequencies during the progress of its POST, so checking for the average signal value won't help you much other than "there's some signal present". It will output 1 kHz when it passes POST; it will output, IIRC, 4, 2, and 1 kHz depending on which step of the POST it's executing.

Failing that, I would suggest you check the voltage rails on the logic board. There is a triple semi-discrete (I mean, there's an opamp, TLC2274 or some such, so not fully discrete I suppose?) regulator providing 3.3, 2.8 and -2.6V. See pic - not mine, but from another user, JFJ. If there's a problem with those rails, the board will, of course, not pass POST.



Edit: you could also try setting the contrast - push "Display", and then the bottommost soft buttons to increase/decrease contrast. It's conceivable that the contrast is so completely out of whack that you don't see anything. Admittedly, it's a bit of a long shot, but hey, if it works, then it works

[al_m]

Hi George, thanks for the pointers. You are of course right about the frequency being opaque just using a meter on the probe comp lugs. At least we know something is going on! I have a frequency counter sitting on top of my scope actually, only they're both inaccessible... Sigh. Anyway just checked and the voltages at the regulators seem ok. 3.3V rail reads slightly low at 3.23V, but the other two are pretty much spot on. I'm in one of those catch 22 situations where i need a scope to really be able to diagnose this scope! Where next? I'm wondering whether the caps on the display board may not be worn out or something like that? There's a lot of small electrolytics on there pushing 20+ years of service...

[al_m]

Oh, the contrast buttons don't work btw. Sadly. That would have made my day!

[TERRA Operative]

I have heard that the capacitors in the PSU in these units can fail.
My TDS210 is still working fine without a recap (ex university scope, so decent hours on it) though so I guess it's pot-luck if you got a good or bad PSU.

I would suggest checking ripple on the PSU output as a sanity check, but you need another scope for that. Do the voltages all look ok from the PSU?


Also, if you don't have it already, here's the service manual:
https://www.tek.com/oscilloscope/tds210-manual/tds200-series-service-manual

[george.b]

At least we know something is going on!

Yup - completely dead it ain't; whether that's a good thing or not remains to be seen, though

3.3V rail reads slightly low at 3.23V

3.23 is most likely fine. I don't think the 3.3V rail is critical.

I'm wondering whether the caps on the display board may not be worn out or something like that? There's a lot of small electrolytics on there pushing 20+ years of service...

Bad caps on the display board was one of the many problems my own TDS220 had. The symptoms were different from just showing nothing, however:



Then again, symptoms could certainly be different from what I experienced.

I'm in one of those catch 22 situations where i need a scope to really be able to diagnose this scope!

Indeed you are. Measuring the frequency of the probe comp output would immediately give you an idea where (not) to focus your troubleshooting. Couldn't you order an el cheapo Arduino thingy online and use it to measure the frequency? Plenty of Arduino frequency meter projects floating around the 'net that you could use. Other than that, I think you will need proper instruments indeed.

Where next?

If not a power supply issue of some sort - TERRA Operative is right, by the way, the capacitors on the PSU may leak, I fixed a friend's TDS210 where, again, among other problems, most of the capacitors there had leaked, so you could check that, but you'd probably be able to smell it when you opened the scope already - that would make it likely to be a logic issue of some sort. A bad ASIC is possible, although I haven't seen it happen yet. This would mean very bad news.
Something that happened on mine is that it wouldn't even give out a probe comp output. I decided to check whether the processor was even running - lo and behold, it wasn't; the /HALT line was asserted. With another scope, I determined that it happened soon after /RESET was deasserted. I made an educated guess that if the processor were reading garbage on the data bus, it would throw an exception, and if it read garbage again when trying to handle the exception, it would halt. The explanation I could come up with is that there should be a bad solder joint on the EEPROM containing the firmware - in my case, a BGA chip - and it was seemingly a correct guess, because reflowing the chip brought the board back to life.
Now, I'm not saying your case is the same - mine wouldn't even output a probe comp signal, so something is different there - but checking for bus activity etc could reveal something. You'll need a scope for that, though.

[al_m]

Hi,

Thanks again for all your help chaps. I decided to order one of those cheap chinese hand-held scopes (FNIRSI 1C15) from eBay. Hopefully it will be good enough to shed some light on what's going on here plus be useful to have around in case i can't get this Tek 210 back on its feet. Will report back once it's arrived.

[al_m]

Hello again. My Fnirsi scope arrived today and first thing i did was measure the probe comp output of the 210. It's giving a ~117khz wave. I'm guessing it's a proper square waveform, but the probe that came with the Fnirsi is absolute rubbish and doesn't appear to have a trimmer anywhere to sort out the compensation so what i see looks a bit like a shark fin. Will have to go digging around inside the unit later and see what other signals i can find.

[TERRA Operative]

Sharkfin shape means you probably have AC input coupling, can you switch to DC input coupling?

[george.b]

Hello again. My Fnirsi scope arrived today and first thing i did was measure the probe comp output of the 210. It's giving a ~117khz wave. I'm guessing it's a proper square waveform, but the probe that came with the Fnirsi is absolute rubbish and doesn't appear to have a trimmer anywhere to sort out the compensation so what i see looks a bit like a shark fin. Will have to go digging around inside the unit later and see what other signals i can find.

Something's not right here. It shouldn't output 117 kHz. Measure the frequency "manually" (using the reticle), see if it makes sense. Try other horizontal ranges.
I wouldn't trust that Fnirsi too much - not just the probes.

[al_m]

Hi both of you, the Fnirsi is set to the following:

1v per div
1us per div
AC coupling
normal trig

This is what it shows me



Sorry for the poor quality iphone photo, but it's all i had to hand. Measurements look right to me if the scope is doing what it should. I will try and get a better probe, but in the meantime i have to work with what's here now...

[al_m]

Ok, a bit further. The power supply rails all read within spec. Ripple doesn't seem to be too bad. The square wave on pin 7 is there and within spec. Main oscillator seems to read ok, comes up 1v pk-pk with a triangle wave in the region of 60.6mhz. Might have a bit of DC offset, is that normal?

Where things get fishy is on the display lines. The four data pins all stuck at 5v no activity even when power cycling the scope. The clock signals are there but all read too slow compared to the expected freqs, and the last one on pin 12 is totally weird. It phases in and out over a long period, like theres some very slow AM on the signal. Really strange...

[andy3055]

Hello community,

after being a passive member of the forum for a while, mostly reading and looking for interesting and helpful stuff...the time has finally come...I need your help

I bought a defective Tektronix TDS210 on eBay. The seller said, the display wouldn't show anything. So i got it for 50€, hoping it would be an easy fix (like pluging back in the powersupply to the main board).
After unboxing and taking the unit apart, I saw that the previous owner has already changed an exploded electrolytic cap. He was even so kind to include the defective one into the package 
Unfordunately the power connector was plugged in correctly, so there vanished my easy fix. So I powered the powersupply up, without connectin anything -> the voltages were within specs. After that I connected the other PCBs and testet it:
The backlight of the display works fine, but the display stays black. So I checked the oscillator and bingo, the output is continuously on 3.3V and no signs of life from the oscillator.
The oscillator is a 60.606 MHz type (see photo) but finding a replacement part seems difficult. I found a possible replacement (see attached datasheet) but these guys cost 15€ plus 18€ shipping on DigiKey.

Here comes my question:
Is it possible to replace the defective oscillator with a 60MHz type? At least for further testing.
Has anyone located in Europe one of these 60.606MHz oscillators at hand and could send me one (to Germany)? This may be cheaper than the DigiKey costs 


Thanks,
Alex

OP: UK RS has a 60 MHz crystal: https://uk.rs-online.com/web/p/crystal-oscillators/4785317/

Not sure if it will work anyway!

[george.b]

Well, it sure does seem like it's outputting a 117kHz wave. Very weird.

The display lines not showing activity is to be expected if it's hanging before the display comes up. Check for activity on the main processor bus. Check whether the processor has a clock (IIRC it's derived from the ASIC). Check whether the ASICs are giving out any signs of life (communication between themselves, communication with the CPU/RAM, etc).

[al_m]

Hi george, yes it does seem pretty weird!

I've done some probing around the exposed chips, which i assume are the processors and some memory. Story gets even stranger from here. There's quite a lot going on but it's hard to tell with the crappy Fnirsi exactly what, as the triggering is horrendous. Anyway, i'm finding a 140khz clock on a lot of pins. Also the 60mhz oscillator looks like it's reaching the National chip ok. Then there is the TI HT244 buffer which i've marked on the picture. I think all the signals on the display pins are present on this chip, which i assume is being fed from the National ADG422 near it. Also, very odd, but the 117khz wave is there, and it comes up as a really sharp square signal.

Honestly, this is a bit beyond my skill level. Without a schematic i'd have to get the shields off and try to reverse engineer it to find where things might be going wrong, but i don't have the time. I wish Tektronix were more forthcoming in the service manual, but it just says 'replace the main board' for nearly any problem i might be facing, so i assume it was really just meant to be used by registered Tek service engineers who could order spare parts.

Really appreciate your time and help so far. Thanks for trying to figure this one out with me..

[george.b]

The 68SEC000 (main CPU) should get a 7.something MHz clock, IIRC from one of the ASICs. It should be present on pin 6 of the CPU. Also check the status of /HALT (9) and /RESET (10).
If the clock is present, and my recollection of the clock coming from one of the ASICs is correct, that means at least that ASIC is working to some degree.

The service manual for the TDS200 series is almost useless it comes nowhere near the quality of older manuals.

It's not like I know totally what I'm doing - I'm looking for clues

[al_m]

Hi, just checked and the CPU does indeed have a 7.58 mhz clock on pin 6. Waveform looks a bit funky, but it's there. Pins 9 & 10 are both high (~3.2V). I've also tracked down those signals that i said in my last post were at the HT244 buffer. Looks like they're coming from the National ADG422. Except the 117khz signal at the probe comp. No idea where that's coming from yet. I'm going to remove my bodge wires sending power direct to the display connector tomorrow and see where that gets things. Hopefully i'll discover something obvious, otherwise the likelihood is this will have to go on ebay as a spares or repairs. Bit of a pity. I was hoping it would be more clear what was wrong with it, but as the power rails seem within spec, it's getting into mystery territory. The fact it's a teeny weeny surface mount double sided board with thousands of vias definitely isn't helping. Tracking down a something like a cracked 2mm long capacitor isnt my idea of fun!

[george.b]

Hi, just checked and the CPU does indeed have a 7.58 mhz clock on pin 6. Waveform looks a bit funky, but it's there. Pins 9 & 10 are both high (~3.2V).

You could also check /BERR (pin 12). It should be high as well. An unresponsive device on the bus (like a faulty ASIC) could cause it to be asserted. Is there activity on the data/address lines, like, say, on lines D0 (pin 61) or A0 (pin 19)?

Except the 117khz signal at the probe comp. No idea where that's coming from yet.

If memory serves, it comes from one of the ASICs. I want to say it's from the same one that controls the display, but I'm not sure. It shouldn't be outputting that, which makes me suspect a bad ASIC. I've never seen this before, and a cursory internet search reveals nothing.

Hopefully i'll discover something obvious, otherwise the likelihood is this will have to go on ebay as a spares or repairs. Bit of a pity.

That would be a pity indeed, as it's a nice little scope, even if it's quite old. Maybe you could find a replacement logic board?

[al_m]

Hi george,

Well, after getting a datasheet for the 68000 series and checking all the pins on the CPU, it looks like it's working, or at least doing something. I realised i've probably been shorting pins out and then the halt signal gets fired or something, as by probing in the vias i can access some of the pins without risk of touching other ones and there's a lot of activity on the D and A pins. I took out the display and probed around the circuit there, it's getting all the signals which are present at the connector on the main board. So, my last step is going to be to remove my bodge wires sending power to the display, do some more checking of signals, and then if i can't find anything really obvious then it's going on ebay, which is a shame, but that's life, i only have limited resources to dedicate to something like this, plus i find it a bit like trying to read war and peace one letter at a time, backwards, and in russian...

[al_m]

The sad end to this tale is that my TDS210 is now on eBay for spares or repairs.

Taking off my bodge wires put it back to where it was when i got it, with no power going to the display connector, and the data lines just held at 5V.

I got a Tek 2225 for a decent price, so will have to make do with that, but it's not nearly as portable!

[blade]

Hi all,

Came across this post while trying to bring back to life an old TDS220, so i thought it would be worthwhile to post here how far i got. I got the scope off eBay, had a damaged display, lose BNC connector, lose rotary encoders and a broken power button. I replaced the screen and fixed everything else and when I powered it on, the display was black and nothing else. So I took the service manual and started from there.

All power signals are ok, but there was no -24V on pin 2 of J302 display connector. I did not want to put a jumper wire to force power, so I did a bit of multimeter probing  and figured out the circuit. I attached it for reference. The signal comes from U310 (ADG422) pin 97. This pin is always low. At this point, I believe that the ASIC is waiting for some OK signal from the processor to enable display power. I manually grounded the base of Q130 to simulate the ASIC enabling -24V line. The display turns white so everything is fine.

Moving my attention the the processor (Motorola MC6800), I checked /RESET, /HALT, /BERR. They are all at 3.3V. By the way, on this board revision, /BERR is directly tied to +3.3V, so I guess it is not used anymore. So the processor is in a running state, it has the a clock signal from the ASIC as well, both chips are fine, I would say. In the MC6800 datasheet, I found that there are 3 pins that output the processor state. These are FC2,FC1 and FC0 (Pin 16, 17 & 18). These pins read 110, which means the processor is in Supervisor program. According to an old scanned book found online, the Supervisor state can control an external memory management device and system software. I believe the processor is trying to read from flash and it can't.

Checking flash memory this is a AM29DL800B in BGA form. Not good since you can't probe BGA pins. Luckily the mainboard has a second footprint U430A for the same chip in  standard 48 pin TSOP. First I reflowed the flash chip thinking the solder must have cracked over the years. No change in the scope status. I then began to check all pins if they have a connection with the address and data bus. The flash is set into byte mode (pin 47 BYTE# is tied to GND), so there are only 7 data pins (DQ0->DQ7) and 20 address pins (A-1,A0->A18) connected. They all have connection with the processor and ASIC. I then checked control pins (WE#, RESET#, RY/BY#, CE#, OE#. The CE# is tied to GND pin 29 on ADG422, meaning the flash chip is always selected. All control pins are connected to MC6800 except one 


[TL;DR]

OE# is not going anywhere. So basically, the processor can not enable the output on the flash to read the data. I suspect a broken PCB trace here. The board is 6 layers and 20+ years old, so it is kind of a given situation. Since I do not have another board of the same type, can someone here with a TDS220/210 that knows or can probe for me where OE# is connected to? I attached a picture where the OE#  is located on the PCB.

[PwrElectronics]

RE: blade....

Thought I could help here but apparently the TDS220 I have access to is older and the board is different...   

I powered this one up and it was working before I opened it up for the pictures.  Don't match yours unfortunately.

[TERRA Operative]

I checked my scope, that pin doesn't seem to go anywhere obvious that I can see with a continuity test all over the PCB, besides 616ohms to ground.