On 01 Jun 2016 14:31:44 -0700, you wrote:

Hi,

So, since we only live once, I went ahead, did a cold start and tried the CCD adjustment as described in the manual, with the PG502. Here is the result: https://goo.gl/photos/wepcezjmfU6uLMH57 https://goo.gl/photos/wepcezjmfU6uLMH57

This looks somewhat better, than than before the adjustments: https://goo.gl/photos/vDnGt2RD5M8LkyZL7 https://goo.gl/photos/vDnGt2RD5M8LkyZL7

I am still dubious about the performance of your PG502 for this. That
looks like a bimodal distribution (two peaks) and is what I would
expect of the jitter from an RC oscillator.

I have to say, it wasn't hard to do, but my eyes are still pogoing after looking at the screen of the aliased display.

I get a headache after a 10s of minutes which is why I have not
experimented with it more.

I did try doing it without the aliased display, and i think the reason for the alias is that on a normal display I get very few samples, so it is really hard to align the CCD sides. On 2ns/div with only two CCD sides on, there are only a handful of dots on the screen and it is virtually impossible to align.

You would want to turn REPET on for this to fill in the aquisition
record. It defaults to off which always struck me as strange. The
display update rate will be slower than for the aliased case but at
least it will be stable.

Having said that I still need to verify if the scope is any better or worse than before, but it was an interesting experience.

Szabolcs

Ok,, I may not have to go that route as I have some more info but first have alook at this short video I made of the prob.
https://www.dropbox.com/s/x5ib2q1yyc5qvju/TDS694c%20CH3%20vs%20CH1%20edge%20trigger-tag%23105206.mp4?dl=0 https://www.dropbox.com/s/x5ib2q1yyc5qvju/TDS694c%20CH3%20vs%20CH1%20edge%20trigger-tag%23105206.mp4?dl=0

Now...I CAN get BOTH PULSE and WIDTH modes of the triggering to work on ALL CHANNELS. ONLY the EDGE mode seems stuck on CH3/4.

So it isn't the trigger chip, it just might be a loose connector from the uC board feeding the ctrl lines governing the acquisition for the edge mode.

Hi,

So, since we only live once, I went ahead, did a cold start and tried the CCD adjustment as described in the manual, with the PG502. Here is the result: https://goo.gl/photos/wepcezjmfU6uLMH57 https://goo.gl/photos/wepcezjmfU6uLMH57


This looks somewhat better, than than before the adjustments: https://goo.gl/photos/vDnGt2RD5M8LkyZL7 https://goo.gl/photos/vDnGt2RD5M8LkyZL7


I have to say, it wasn't hard to do, but my eyes are still pogoing after looking at the screen of the aliased display.
I did try doing it without the aliased display, and i think the reason for the alias is that on a normal display I get very few samples, so it is really hard to align the CCD sides. On 2ns/div with only two CCD sides on, there are only a handful of dots on the screen and it is virtually impossible to align.

Having said that I still need to verify if the scope is any better or worse than before, but it was an interesting experience.


Szabolcs

You might try pushing on various areas of the board and various parts
with an insulated tool while it is operating to see if there is a
particular spot which is mechanically sensitive.

I would also try reflowing all of the solder joints. If that did not
work, I might cut the top off of the hybrid. I do not think you have
anything to lose at this point.

I am not sure which other oscilloscopes shared the same vertical
amplifier design but the 464 and 468 are closest followed by the 465B.

I do not know why they use the aliased display but I also have not
compared to two methods in detail yet.

The non-aliased display uses multiple triggers so any trigger jitter
is added to the display. The aliased display acquires an entire
record from one trigger (or no triggers) and multiple cycles of the
source so the long term source jitter is important and I would not
trust the RC timing of the PG502 for that; I expect it is
significantly worse than the LC timing of the SG503. RC oscillators
have lower Q and higher phase noise and higher jitter than LC
oscillators.

As far as the procedure, just display the non-aliased output using
your PG502 and the fastest time/div setting for a good display and
make the CCD timing adjustments as described in the manual.

I think it was mentioned here that Tektronix had a special employee
who had the knack for doing the CCD timing adjustments as described in
the service manual. She was just better at it than anybody else.

Hi- I used hot air from top of PCB.
If I had any issues, I would have used my preheater to heat from bottom - but turns out it worked fine without use of preheater.
James

On 01 Jun 2016 09:17:18 -0700, you wrote:

An Armature Radio Swap meet would have the items you are looking for
inexpensively. I make my own patch cables from connectors and cable
like RG-400.

I was planning on finding some RG-58 and getting some connectors and make a couple cables

RG-58 works fine for this although it may be marginal at 200 MHz and
above. An Armature Radio swap meet will usually have cheap 50 ohm
patch cables also.

Probe to BNC adapters can be found on Ebay. Most of the inexpensive
Chinese probes come with them.

Maybe you can point me in the right directions cause I have search Ebay and found a couple for $30 or more for a single unit.

You might not have recognized them. I have one official Tektronix BNC
to probe tip adapter and a bunch of the cheap Chinese ones which work
just as well.

http://www.ebay.com/itm/Probe-Tip-Adapter-allows-connecting-probe-to-scope-port-DiLette-DL-PA009-BNC-/252406586228?hash=item3ac49adb74:g:KwsAAOSwepZXSz3p

I would leave the various tantalum electrolytic capacitors alone.

I pulled the list off the service manual and looking at my scope.. I was sure which one where Tantalum Electrolytic Caps or high voltage ceramic cause the manual list them as ELCTLT which I thought meant they where all the same

Unfortunately the parts list in the service manual is not specific on
this issue. Sometimes you can tell based on the tolerance or the
manufacturer's part number; if tolerance is worse than 20%, then it is
aluminum.

I examined a couple of 2235 series main boards that I have and *all*
of the aluminum electrolytic capacitors are radial style parts (both
leads on one side) which stick up. Any axial style capacitor (one
lead on each end) was either ceramic or tantalum.

An LM317 is fine as far as noise goes; they are relatively noisy but
quiet on the scale of oscilloscope input noise unless you are using
something like a 7A22. At higher output voltages, a bypass capacitor
from the adjustment pin to ground will lower the output noise as if
the output was set to 1.25 volt.

The 10:1 probe is a problem as you identified. When I used a 50 ohm
x10 attenuator with 50 ohm load to lower the output of my PS503 to 0.2
volts because otherwise it was difficult to set exactly, the parallel
25 ohm output impedance allowed for an accurate measurement with a 10
megohm input DC voltmeter. Had I tried the same trick with a x10
probe, the parallel output resistance of about 1 megohm would have
caused considerable error although using a high input resistance
voltmeter might work.

As far as an easy way to get a stable 0.2 volt output from the 317,
instead of modifying the regulator circuit for a 0 to X volt output,
add a low impedance resistive divider to the output. This would not
work for a regulated supply application but is no problem here.

Alex, previously you measured 14.4 V between pins 9 and 4 of the time-delay tube 6NO45. Then about 20 V -14.4 V = 5.6 V should remain between pin 4 and ground, so 5.6 V across the coil of the relay. Now you mention 0 V across the coil. Could you check the +20V supply?
The weird thing is that the time-delay tube worked fine in your other 'scope. Could there be something wrong in the chassis socket for the tube? Note that internally inside the tube pins 3 and 4 are connected and pins 8 and 9 are connected.

Albert


---In TekScopes@yahoogroups.com, <alex_brinister@...> wrote :

From a visual inspection, the relay looks fine to me. The resistance of the coil seems to be correct, 150 Ohm. This is the same (as expected) as my working 585A. However, the relay doesn't click. The voltage across the coil of the relay is 0 V. To be clear, the relay never clicks. The warm up period never ends. After ~30 seconds, my other 585A clicks and the voltage across the coil is 18.4 V, which is what should happen.

This would imply that K600, the warm up tube, is defective...

To make sure K601, the relay, is working - you could try to short pin 9 and 4 and see if the relay clicks. You should only get one click that won't "unclick" - one you short 9 and 4, it will energize the relay and then it will latch on until you power down.

But before you do that, make sure that you do get voltage across pin 1 and 6 of K600, and also make sure all wires are connected correctly. Also the resistance across pins 1 and 6 should not be open. Perhaps the heater in K600 burned out?

If it comes down to replacing K600 I wonder how many people would mind if it gets swapped with an RC time delay, op amp, and some components... That would seem wrong and lose the classic feel of the scope? No?

Hello all, Sorry for not replying earlier, did not have time to look at the scope again until today.


Yes, I think I did confuse K600 (the tube) with K601 (the relay).


From a visual inspection, the relay looks fine to me. The resistance of the coil seems to be correct, 150 Ohm. This is the same (as expected) as my working 585A. However, the relay doesn't click. The voltage across the coil of the relay is 0 V. To be clear, the relay never clicks. The warm up period never ends. After ~30 seconds, my other 585A clicks and the voltage across the coil is 18.4 V, which is what should happen.


The plugin... I bought the Tektronix-made 013-0055, which is longer than I thought. I took the connector out of the plugin thinking I would make a longer flexible extender. I definitely should have waited. The 013 is long enough. But the plugin I cannibalized had three missing transistors, which is what I was concerned about. One I found replacements. The next, a 2N212, I could not. I used the site alltransistors.com to find a replacement but it didn't list any and an NOS one from eBay is ~60 dollars. I could attempt to solder the connector back on (I have other 82s that I could use for reference or the manual) if I can find a similar transistor to the 2N212. The last trasistor is a TA2159, which I found on tek-parts.com for cheap (5 bucks).


Thanks,
Alex Brinister

Hi David,

I checked connections to pins 1 and 5 and they seem to be fine. What is interesting, though, is that after I reseated the hybrid, the scope permanently went into the "second mode" where it has 1cm of deflection; now its behavior doesn't change after the scope warms up. I guess there _was_ an intermittent connection, but it was not the only problem.

This being said, it looks like the IC is indeed faulty. I'll try to find a replacement for reasonable price.

Thank you for your help and expertise!

Yes the +10V is fine, I recall measuring it at 10.01V when I was dealing with the false overload issue.

I was just commenting on the my situation with poor external equipment to do the calibration though good enough to trick the scope to believe that I'm not lying to it :)

Rob,

Well it's a good sign that dots appeared on the screen, at least it was oscillating! But not so good when they faded away :(

Anyway maybe I'm not reading something correctly here. Due to the back to back diodes CR4104 and CR4010 (well, more of CR4010), the -8 V supply cannot reach the base of Q4008 and thus cannot reach CR4005.

The circuit must start oscillating when CR4005's cathode is positive (CR4005 is there to protect the filter capacitor C4003 I suspect). Initial conditions will get Q4008 to pull that wire up to +5(minus saturation voltage) current limited by R4002. That it will flow through pin 6-7 of the transformer and turn on Q14009. This will cause induction to reverse polarity of winding 6-7 forcing pin 7 to go negative - and this is where the negative voltage comes from. This negative voltage also will force Q14009 to shut off. This will then induce the opposite voltage in the 6-7 winding which will cause pin 6 to raise up again (which is bounded by the voltage at pin 7 - which is controlled by Q4008 and cathode of CR4005.

The feedback circuitry basically tells the oscillator to shut off if the voltage is too negative (this is connected to the ~ -2000V cathode of the tube) as it will turn off Q4228 and Q4233. No more base current into Q4008 will shut it off too, and the oscillation will keep on pumping negative voltage into C4012 until it gets too negative and oscillation stops as the base of Q14009 can no longer get high enough to turn the transistor on due to the negative offset.

At least that's what I think how this circuit should work.

This means if CR4005 cathode is positive, the oscillator should be oscillating as the transistor should turn on at some point. However if the voltage is too high, it will pull Q14009 into permanent saturation and killing it. The voltage will be positive there if oscillation is not happening or if the -2000V can't get negative enough. If it's stuck there for too long, bye bye fuse and/or Q14009.

Now is where a storage oscilloscope would be good, or at least a fast (analog?) multimeter that you can see the meter peg negative... What's the voltage over time at that cathode-CR4005 node? Did it ever go negative? Yes=secondary overload No=primary problem.

What is the waveform at TP4009/base of Q14009 like - is it hitting +0.7V too - if it's staying above 0.7V, Q14009 would be hard-on and will fry itself. It still seems that Q4228 never turns off due to overload on one of the outputs which is still an issue.

At least the CRT is seemingly still good as it showed dots at least for a little while.

SMD electrolytics in the control panel. If you've replaced them you may need to clean the front panel PCB as best you can to remove any dirty shorts.

Look for the group photo album I just uploaded of the 420 that I cleaned and rehabbed, it has a shot of the front panel PCB where you can clearly see the SMD caps next to the speaker.



Assuming a leak would have followed the gravity vector, C401 is your culprit as it looks to be above the circuit for the speaker (the photo is upside down). If the pollution was able to creep upwards, your culprit would C410. You may never know, so clean clean clean and replace all of the caps.

These units get a bad rap for the PSU and the massive amount of SMD caps to clean up but they're nice little machines if you take the time to clean them up.




---In TekScopes@yahoogroups.com, <tigrol.lechat@...> wrote :

Hello,

Have you also replaced and cleaned-up SMD electrolytics on front panel?

Paul

2016-06-01 14:47 GMT+02:00 marc.zoellin@... mailto:marc.zoellin@... [TekScopes] <
TekScopes@yahoogroups.com mailto:TekScopes@yahoogroups.com>:

Thanks,

What would be the procedure for that? BTW, what is the reason for using an aliased display? I'm sure there is one.


Szabolcs

Hi,

Have you checked the 10V reference voltage in the scope? And the other voltages for value and ripple?
Also, check the voltage measurements on both channels and at, may be one of the attenuators are stuck in the scope. Do you hear both clicks when going over the range of the V/div knobs?


Szabolcs

Hello,

From a cold start the display will slowly zoom for about 30min, then zooms
a lot faster loosing brightness and linearity before completely loosing the
display.

Since I have replaced U438 (LM317) the whole PS is brought down when warm
rather than only loosing display.

Although I have made no measurements with CRT and driver board out of the
scope yet, I'm starting to question the FBT, the red potting appears quite
aged and wrinkled, do these have known issues?


Paul

David and Group,
Thanks for the response. The restorer diodes (A4CR4122 and A4CR4123) are P/N FDH5004. Where can I buy these? I want to replace them. Also is there are good source for the HV caps? It looks like the HV caps are: C4016, C4020, C4025, and C4126.
Regards,
Rob 

Sent from Yahoo Mail. Get the app

---In TekScopes@yahoogroups.com, <davidwhess@...> wrote :

>did not go through your list in detail; I just showed some examples.

I spent some time going through the service manual and that was more or less what I found

I would leave the high voltage capacitors alone. I have never heard

of them failing in a 22xx oscilloscope.
Ok Sound reasonable

An Armature Radio Swap meet would have the items you are looking for

>inexpensively. I make my own patch cables from connectors and cable
>like RG-400.

I was planning on finding some RG-58 and getting some connectors and make a couple cables

>Probe to BNC adapters can be found on Ebay. Most of the inexpensive
>Chinese probes come with them.

Maybe you can point me in the right directions cause I have search Ebay and found a couple for $30 or more for a single unit ..

>The parts I would change in rough order of importance:

>1. Aluminum electrolytic capacitors - all of them.
>2. X and Y safety capacitors which are located at the AC line input.
>Film safety capacitors can be used as replacements if you do not like
>the idea of paper capacitors.
>3. The focus resistor chain. I would use HV25 and HV35 style high
>voltage film resistors to replace them.
>4. Check the condition of the vertical CRT amplifier collector load
>resistors. Replace them if they are burnt or cracked.

Thanks for the help and will update the list


>I would leave the various tantalum electrolytic capacitors alone.

I pulled the list off the service manual and looking at my scope.. I was sure which one where Tantalum Electrolytic Caps or high voltage ceramic cause the manual list them as ELCTLT which I thought meant they where all the same ..