Re: 2465 DVS Repair - was; Re: [TekScopes] Wanted; Person to repair Tektronix 2465 DVS

Chuck Harris

The power inside of the 2465 supplies is pretty dirty from
an RF point of view. It is pretty quiet from a 2x60Hz point
of view.

How can you tell the difference?

The RF noise is composed of sliver thin spikes that ride on
the top and bottom of the supply's DC waveform. They look like
a fuzzy caterpillar.

The 2x60Hz ripple will stand still when the scope is triggered
using the line mode.

Tek recommends you use the scope in bandwidth limit mode for
making these measurements.

The other failure that can happen to cause the Test05 Fail XX
failures is in the MUX and Sample and hold gates that the A5
board uses to feed the threshold signals to the trigger hybrid.

Typically one of three things happens (In approximately the
order of likelihood):

The MUX has an output that gets stuck high, or low (the message
gives a clue), or,

The sample and hold op amp gets stuck high or low, or,
The sample and hold capacitor gets leaky.

Trace back the TLA and TLB signals from the trigger hybrid to
the two sample and hold OPAMPS, and from there to the MUX.

And, it is conceivable that the AC Line signal that is supposed
to go from the power supply to the CHN5 input to the trigger hybrid
could be missing, or stuck... TP46 is your friend.

-Chuck Harris

jhalbrecht wrote:

O.K. tried without success to improve the ripple signal on the DC test points. I
don't have a dedicated ground on the Rigol nor on an old 15mhz B&K I aoso tried. I
tried with the banana clip between the 2464 ground jack and the ground of the test
signal on the rigol and on the outside of a BNC connector on CH2 of the Rigol. I
tried it like that with and without the addition of the probe ground too. Nothing got
me near to what I thought ripple might look like.

Is there any change the DC is that dirty?

I also tried swapping out the U300 trigger hybrid from the 2445 switch and knob
donor. Nope, same error Test 05 Fail 42

- Jeff

On 9/14/2019 4:59 AM, Chuck Harris wrote:
That won't do it!

You must connect a big fat banana plug terminated cable in the
2465's ground jack (near the calibrator) to the equivalent ground
on the front panel of the scope that is measuring the ripple.

Anything less will allow the ground loop current to pass through
the probe's shield, and the current through the probe's shield will
generate tens of millivolts (I2R) of unnecessary ground loop noise
in your signal.

We are dealing with a maximum of a few mv of signal here.

Strap the two scopes' grounds together!

-Chuck Harris

jhalbrecht wrote:
:-) yup. Connected the short ground clip from my probe to the solder ground lug on
the chassis / bnc of ch 1. Also set 20mhz bw filter too.

- Jeff

On 9/13/2019 9:11 PM, Chuck Harris wrote:
While doing the ripple measurement, you must connect the
grounds between the 2465 and the testing scope.

-Chuck Harris

jhalbrecht wrote:
Before trying the hybrid swap I got a bit more familiar with my 2465 by testing the
voltages on J119 and comparing to table 5-1 in the adjustment procedure.

I used a Keithley 2000 Multimeter and a Rigol DS1054z oscilloscope with the extra
options enabled.

The measurements looked solid and stable well withing the published tolerances.

However I'm not confident on my procedures for measuring ripple. It seemed more
stray noise basically at 120Hz I tried with the Keithley on AC and the Rigol
with AC
coupling. All the ripple appeared well below 100mv but for instance on pin 6, 12
5 that exceeds the 15mv spec as well as the 10mv spec for pin 1.

Before actually physically swapping the trigger hybrid are there any other test
suggestions to try first?

- Jeff

On 9/12/2019 11:11 AM, Chuck Harris wrote:
Hi Jeff,

Usually, when I see Test05, Fail XX, and it isn't the A5 board
of a late model 2465B, it is due to a bad trigger hybrid. Try
removing one from one of your 2445's, and put it into your DVS.
I would bet it will fix the problem.

Note: Seat the chip by hand, making sure it goes all the way
to the PCB. Then while holding it, put nuts on diagonal corners,
and snug them finger tight. Then put on the other two nuts.

Tightness doesn't improve the contacts, it just breaks the board
and studs. No more than 1/8T beyond contact. Just enough to
make the lock washer grab, no more.

Do measure the power supply's voltages and ripple. Don't tweak
any adjustments, unless you want to do a complete recalibration.

The capacitors are almost certainly dead by now... especially
if the fan is stopped, or sluggish from lack of lubrication.

-Chuck Harris

jhalbrecht wrote:
I got the paddle switch out of the donor without incident.

Chuck and group,

Perhaps I could check some voltages on the 2465 dvs (original / first model) A5
next in an attempt to solve the "Test 05 Fail 42' triggering problem before
into replacing the paddle switch into the 2465 switch board assembly.

- Jeff

Join to automatically receive all group messages.