开云体育

On Wed, Apr 30, 2025 at 11:36?AM Luca via groups.io <la_whg=
outlook.com@groups.io> wrote:

Thank you all for the tips, I made a mistake, and didn't check C42 at
all...I missed it completely. I wasn't completely dead, but halfway there
which is why the problem only manifested itself with the power supply fully
loaded. With C42 and C61 replaced (C61 was fine, didn't make a difference)
the ripple is now within spec on all rails (10mV or less).

Awesome - he who never makes a mistake makes nothing :). Next time I bet
you'll remember to poke a probe at the unregulated supply before you drill
into the regulated side.

Hi all

Thank you all for the tips, I made a mistake, and didn't check C42 at all...I missed it completely. I wasn't completely dead, but halfway there which is why the problem only manifested itself with the power supply fully loaded. With C42 and C61 replaced (C61 was fine, didn't make a difference) the ripple is now within spec on all rails (10mV or less).

Two of the metal can transistors still get hot, but at least I can keep my finger on it for few seconds. The voltages are stable now. I have the color bars back, but my "Oven temp normal" still doesn't come on.

There are +10V on one side of the lamp, but my TP1100 is still at 3.98V

/g/TekScopes/photo/296482/3911053

There is a note in the manual, maybe it's also worth a try to check.

Page 4-9
CHANGE TO:
Crystal and Oecillator Oven Assemblies

Because of the critical nature of the oven sub assembly, the entire assembly of the Oscillator Board, A10, and the Subcarrier Board, All, should be replaced as a unit if the crystal or oven fails. If crystal or oven replacement is attempted, the following pOints should be observed:

1. Silicone grease must be present between the temperature stabilizer and the wrap-around housing, Q1031, R1031, and R1035.

2. When reassembled, the oscillator board must not touch the wrap-around housing or either of the oven cover.

3. The crystal should not contact the temperature stabilizer other than at the welded flange around the crystal base.

4. When plugging the oven back onto the Subcarrier Output board, fit the oven, the two pieces of insulation, and the cover together; then attach to the Subcarrier Output board as a unit. This allows the spade lugs on the cover to align the Berg socket over the 6 square pins on the Subcarrier Output board.

5. To determine the oven operating temperature, a Leeds and Northrup 8693-2 Temperature Potentiometer with small thermocouple, or equivalent may be used. The thermocouple junction should be clamped between the temperature stabilizer and the crystal retainer. R1112, the branch of the temperature-sensing divider network, may be changed to move the operating temperature to 85°C if
component aging has seriously displaced the temperature equilibrium point. Normal operation will not be seriously impaired as long as the oven operates at a constant temperature at any point between 80°C and 90°C.

1 - if there was any grease it must have been dried up long time ago
2 - there is foam that is disintegrating between the inner oven and outer housing so probably also worth checking

5 - is probably worth checking. My multimeter came with a thermocouple or what do you think shall I check next?

On Tue, Apr 29, 2025 at 12:08?PM ed breya via groups.io <eb=
telight.com@groups.io> wrote:

I think Siggi has called it right if the DCV and ripple on the raw feed to
the 3.6V regulator has not already been confirmed OK or fixed. I thought
though that all the filter caps had been replaced earlier. If not, this
could explain it all. If the C42 has been replaced yet still shows
excessive ripple, there may be an open rectifier ahead of it, which would
make it half-wave with much bigger ripple at line frequency rather than 2f.

That's a good point. Looking at the oscillogram (
/g/TekScopes/photo/296482/3910865), it looks to me that
the period is 10ms, which is double 50Hz?
The only way I can see the signal flat-topping like that is if there's no
(or very little) current drawn from the unregulated rail. Though, as we
don't know the DC level, maybe it's not flat-topping at 0V.

All this being said, this current limiting arrangement seems a little hokey
to me as it's limiting the output of the regulator, while measuring the
combined currents from the rectifier and the unregulated output?

I think Siggi has called it right if the DCV and ripple on the raw feed to the 3.6V regulator has not already been confirmed OK or fixed. I thought though that all the filter caps had been replaced earlier. If not, this could explain it all. If the C42 has been replaced yet still shows excessive ripple, there may be an open rectifier ahead of it, which would make it half-wave with much bigger ripple at line frequency rather than 2f.

Ed

Ok, I'm going to bet my $0.02 that C42 is simply dead.
Have you measured the unregulated supply? I'd bet that the unregulated
voltage is dropping below 3.6V + the saturation voltage of Q55.
At that point Q850 is going to try and hold up the supply by feeding the
10V supply through the BE junction of Q55 for all it's worth.
This is no good for either transistor :).

Good luck,
Siggi

On Tue, Apr 29, 2025 at 9:51?AM Sigur?ur ?sgeirsson via groups.io <siggi=
sort.is@groups.io> wrote:

On Tue, Apr 29, 2025 at 9:10?AM Luca via groups.io <la_whg=
outlook.com@groups.io> wrote:

While (apparently) overloaded, it's according to the attached picture

/g/TekScopes/photo/296482/3910865

You need to measure this DC coupled, not AC coupled as the P2P really

says nothing about what's happening here.

On Tue, Apr 29, 2025 at 9:10?AM Luca via groups.io <la_whg=
outlook.com@groups.io> wrote:

While (apparently) overloaded, it's according to the attached picture

/g/TekScopes/photo/296482/3910865

You need to measure this DC coupled, not AC coupled as the P2P really

says nothing about what's happening here.

Ok, for checking Q55 at its real load parameters I have to find a way. I could take it out but as far as transistor testers go, all I have is the cheap multi component tester.

Regarding the ripple, without a load it's correct.

/g/TekScopes/photo/296482/3910866

The voltage drop across R841 is about 0.17 V

While (apparently) overloaded, it's according to the attached picture

/g/TekScopes/photo/296482/3910865

The voltage drop across R841 is about 0.4 V

I did measure R841 and it is out of spec. It's a 0.2 Ohms at 10% tolerance (according to what's written on it) and it currently reads 0.275 Ohms with my highest precision meter. I'm going to check the rest of the resistors too just in case, and try to find out if any drifted resistors are messing up the power supply.

Sorry that I haven't been following this thread too closely, so forgive me if this has already been discussed, but you may need to retest that transistor's beta at collector currents closer to where the transistor is supposed to operate. Beta is not a constant, so a measurement made under some random condition doesn't tell you a whole lot (other than "this transistor sometimes can transist"). Beta is very much a function of current, and unhealthy devices may check ok in some range but not in others. If you are persistently unable to figure out what's going on, you may wish to rig up something to determine what the beta actually is. If taking the transistor out is too much trouble, you'll need to figure out from other measurements what the actual beta is.

Good luck!

Tom

--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
420 Via Palou Mall
Stanford University
Stanford, CA 94305-4070

Something's not right here. If Q55 has plenty of gain, then there's no reason for Q850 to run hot. Are you absolutely sure Q55 checks out OK? I wouldn't necessarily trust the measurement without knowing more about the tester and setup.

I assume you've been using a scope to measure ripple, so you would have noticed if there's any high frequency oscillation that can mess things up. If Q55 is definitely good, then I'd focus on Q850 now and figure out where all the currents are going and how big to account for it getting so hot.

Ed

Luca,

The current is determined by the voltage drop across R841, in the 3.6V case. Ohm's law V/R (ohms)=I. This is the limit (600-650mA), not what the load draws. Measure the voltage across these resistors in the current limit parts to determine the load current. If a supply overloads, usually tantalums are the first thing to check as these will go bad (short).

Test Q55's hfe out of circuit. While it is out of circuit, test the transistor for leakage/open. It would be best to get a new one and any other spares you want. Be sure to put on new compound when you replace the transistor(s), MX-4 for example.

Test ripple while under load.

Mark

Hi

Thanks for all your suggestions. Q55's hfE value is around 580 on those china testers (in circuit).

Check Q850, 2N2192A, and get a spare or several in case. Getting that hot could have changed the parameters even if checks good. the 3.6V supply is limited at 600-650mA by R841. At this point or higher, the supply will shut down by turning on Q840.

Mark, how did you determine the current rating? I'm not an EE just a hobbyist but would like to learn about this because now at least I can try to measure if the load of all modules connected at once exceeds this value.

I did replace the capacitor (the old was slightly tired but not broken), and without a load the power supply ripple is within specifications.

I should have a spare for the 2N2192A *somewhere* but if I can't find it I could try it with the suggested part

Considering the circuit, there is more than enough voltage in the base circuit to use a darlington transistor in place of the 2N2192A.? The circuit should perform normally with a slightly elevated?voltage at the base of Q850.? There are a number of high current NPN TO-92 darlington transistors on Mouser.? You may have a few of the older variations in your junk drawer,?just make sure the IC will handle about 100mA.? It would be worth a try to see if the issue persists. - Ferrous

Luca,

Q55, as Ed B. said, is suspect. Use an hfe meter to test yours. I have seen the originals too low, beta <10. The beta would be nominal 110 at the current range the load wants by the chart for a 2N3055. Make sure to get a replacement from a reliable source. There are a lot of counterfeits out there - especially on ePay. The old phrase, you get what you pay for, is true. You can check the other series pass transistors to confirm their beta.

Check Q850, 2N2192A, and get a spare or several in case. Getting that hot could have changed the parameters even if checks good. the 3.6V supply is limited at 600-650mA by R841. At this point or higher, the supply will shut down by turning on Q840.

It is possible the ripple is high on that supply. A 594-MAL205655103E3, Mouser p/n, could be used as a replacement. See if the decoupling, 100mfd, on the output of the 3.6V supply is good. You may have already replaced this.

Mark

Check Q55's hfe around the load range. If it's not enough gain then Q850 may be working very hard and also allowing the raw supply line ripple to show at the 3.6 RTL supply when Q850 hits its intrinsic current limit.

Ed

Hi

I'm still looking for why my 3.6V supply has issues.

The current draw of the boards doesn't seem excessive (though I don't know what's their normal power consumption - Bar Timing board draws almost half an Amp). Q850 gets scorching hot.

/g/TekScopes/photo/296482/3910379

Any ideas?

I do get a picture, though sometimes I loose the colors and the crosshatch pattern's vertical lines aren't perfect

The capacitors are fine... I did change the ones on the PSU board, but left the large cans be.

However, there was an open diode in the bridge rectifier circuit which was the cause of the ripple on the 10V rail

I also made a mistake and do have all the colors, it was a setting on the monitor.

Now there is still ripple on the 3.6V line which appears to be caused by the 3.6V regulation going into current limiting. It looks like something on the A3 field timing board, though nothing appars to get hot. The A3 board is quite complex....

/g/TekScopes/photo/296482/3889932

Hi all

I have dialed in the voltages, which brought back most of the logic that generates the patterns, but there is still something wrong with the heater circuit.

It generates a picture now, though it's missing two colours and has artifacts in it: /g/TekScopes/photo/296482/3889912

Unfortunately Mark is right, the unit needs recapping because ripple is all over the place and out of spec on all rails (10 V rail has 2V of ripple - yikes)

Hi all

Thanks for your inputs... yes it's mostly for preservation reasons... how many people or institutions are left with a functional Tektronix 145? I initially picked it up for spares but it's so well made it would be sad to trash it if repair is possible. If I can't get the crystal oven to work or other BER situation I can still gut it.

I do have old PAL televisions to repair/restore so it could also theoretically serve a purpose.

I'm not sure replacing the resistor will help with anything.. it reads within spec, is °C 130~ too hot for a larger resistor? I checked all the socketed transistors and they're OK. The thermistor is not open, though I'm not sure about its value (I should probably remove it out of circuit to check). I will have a look into the power supply section and try to get it within spec before I continue.

After 30 mins of waiting, the TP1100 voltage stopped at 3.95 V (the manual says it should come on before 20 mins - the recommended warmup time). Maybe it does need the PSU section dialed in.

Luca,

R1156 is too hot. Replace with a 5W. Look around the TP1100 for bad parts. The thermistor could be open. Dirty pins/sockets could be the problem. Check the ripple in the power supply. At the age of this unit, recapping may be necessary.

Mark

To fix PAL standard televisions?

Yes, people are into preserving antique televisions and equipment.







Now if 145 PAL does not refer to test equipment for the non-NTSC standard televisions them I am out.

Frank DuVal