[Techtoolslist] Help with BK 545 circuit analyzer

[Chris Rinewalt]

To update, I have figured out the issue, but ultimately need help with the
solution. Instead of going back and forth looking at ICs, I just decided to
probe around and see if anything jumped out. I found that when I was near
the 1kV LCD screen power wires (red and black pair running from 10 to 5 in
this photo with only clear plastic tubing covering them,
http://imgur.com/QyV5bBT ) there was a lot of noise. I unplugged the wires
and found that I could easily adjust the 60Hz frequency pot through a full
range and it was steady at all frequencies. There was still a but of noise,
which I am guessing is either coming from either the kV transformer (small
daughterboard north of the LCD power wires in the photo) and/or the other
two transformers, but the vast majority was due the power wire. So I
ordered a combination EMI filter/heat shrink tube and installed it. I
didn't shrink the tube as I was worried the heat may damage the nearby LCD,
so I just fastened it with a zip tie. I found that there was still an issue
with hitting 60Hz exactly; it was fine at 59 Hz then again at 63Hz (whereas
previously it would steady at 75Hz or so). Oddly, the EMI tubing indicated
it should be grounded, but when I hooked up the ground the frequency
shifted down about 4 Hz and the residual noise was much greater and moved
through the waveform quicker. The specs for the EMI tubing indicate it
works at above 10KHz, so I'm glad it did anything at all.

Does anyone have an idea of what I can use to block the remaining
interference? I've read up a bit on mu-metal but that seems only to
redirect the field, whereas I need something to cover the wires entirely,
shielding all other components inside the unit. Becuase of the tight
quarters, there's not really any way I can fit a cage around the analog
board (the upper board in the photo which the interference effects most),
nor over the 1kV tansformer. I also can't really re-route the LCD power
wire as it connects inside the LCD assembly and I'd rather not open it up,
plus I'm sure that anywhere I put it inside the box, there will be noise.
Unless something has been modified in this unit, I'm surprised this design
made it past the final engineering QC. It is usable as is (and improved a
bit with the EMI tubing), but it should have been evident that the noise
from the 1kV transformer was bleeing through when setting it up at the
factory - unless they also removed the power when dialing it in.

Chris

On Sat, Jun 17, 2017 at 1:50 PM, Chris Rinewalt <chris.rinewalt at gmail.com>
wrote:

> Sorry, this is a bit long....
>
> I picked up a BK 545 circuit analyzer that appeared to be operating OK,
> but did have issues. For instance, upon first power up instead of
> displaying a horizontal line, it had an asymmetrical oval shape distorted
> along the neg Y-axis. It would display somewhat correct V-I curves for
> resistors and diodes, but rather than a single line, signatures were ovoid
> an distorted. Oddly, when I turned it on the next day, this issue was gone,
> but there was a fair bit of noise along the curves.
>
> I am still searching for a manual for the unit, but I was able to get the
> schematic and service sheet from BK, so I tried to calibrate it. It was a
> royal pain and some of the procedures I could not get to work properly to
> meet the indicated specs (more on this below), but I have it operating
> fairly well with two issues that I think are related.
>
> First, the 60Hz sampling frequency cannot be adjusted to spec. I followed
> the procedure and it just bounces all over the place (though if I engage
> the low pass filter on the frequency counter it does settle down). To get
> the clock, a TLC555 is used, with a fixed 20K Ohm resistor (R141) and a 50K
> ohm pot (VR105) between the output and threshold (http://imgur.com/2Ez8C6N).
> At 60HZ, F1 and F2 are lo, so it's just the resistor, pot and 470 pF cap
> (C105) in circuit with the TLC555. To attain 60HZ, VR105 is adjusted to
> attain an approximately 30.72KHz square wave from the output. I can adjust
> the pot just fine, but it only outputs an signal of approximtely 30.2Khz
> which gives an frequency of about 58.2Hz. Not bad, but the odd thing is
> adjusting the pot one way has little to no effect while adjusting the other
> way introduces a lot of noise into Q, which causes a noisy frequency output
> that bounces around. It's odd, the signal is fine, turning the pot makes a
> little difference for some travel and then at a certain point the square
> wave gets all out of whack and the output sampling frequency bounces all
> over the place. This happens up to about 48KHz or so (I forget the exact
> freq) from Q when the square wave clears up again, causing the output
> frequency to be about 75+ Hz, too far off. So, why is it that the Q output
> of the 555 is 'clean' for only two small portions of the pots travel,
> neither of which align with the required 59-61Hz called for in the service
> sheet?
>
> Then there is a second issue with the 60Hz signal. As I stated above, I
> had a heckuva time getting this thing spec'd. I came to realize that the 4
> other pots in other circuit sections used in the calibration are pretty
> much dependent on  one another. The noise @ 60Hz was originally so bad I
> figured it had to be an op-amp (TL084) or one of the other CMOS ICs. So, I
> pulled and tested every IC in that part of the circuit (and then some),
> along with the 50K pot. All tested fine, but I replaced them anyway. And
> still there is noise. Even if I adjust VR105 to reach a stable 30.2KHz from
> the 555, getting ~ 58Hz sampling freq, it still has so much noise that my
> counter can't lock onto it without a low pass filter as noted above. When I
> switch to one of the other sampling frequencies (400Hz, 1Khz, 2KHz) there
> are no such issues.
>
> So, here is where I am stuck. I did read somewhere that ceramic disc caps
> should not be used in a 555 astable oscillator circuit, but I have to
> assume BK engineers knew what they were doing and that is not an issue. I
> did test the 22uF electrolytic (C104) on the +5v line and it is OK, but I
> don't have anything to test the 470pF disc cap. I do know it is not
> shorted. R141 is in spec (19.7K). There is some noise on the triangle wave
> on pins 2/6 of the 555, which clear up when any sampling frequency other
> than 60Hz is selected. I replaced the LS390 that Q feeds into but still no
> change. So other than that what else in this circuit could be affecting it?
>