BK Precision 2160A repair

[ZaUcY]

I also just realized that R419 is connected to the base of Q407.  So because I was getting a bad reading for R419 when still connected to the base of Q407 in the circuit, might also suggest that Q407 is bad.

Still not sure why the hFE of Q409 was 30% higher than Q408, the same transistors.

[tautech]

I also just realized that R419 is connected to the base of Q407.  So because I was getting a bade reading for R419 when still connected to the base of Q407 in the circuit, might also suggest that Q407 is bad.

Still not sure why the hFE of Q409 was 30% higher than Q408, the same transistors.

I would be tempted to replace all 4 transistors in the pre-amp stage.
If your tests are with removed devices, things are up the creek. 

The device #'s have been shortened as is often the case and I usually put "2S" in front of the # for it to make sense.
But do your homework re datasheets to ensure you get the correct devices.

[ZaUcY]

I would be tempted to replace all 4 transistors in the pre-amp stage.
If your tests are with removed devices, things are up the creek. 

Do you mean the final amp stage?  that is where Q407 - Q410 are from.  If that is not what you mean, can clarify please.

Mostly voltage is the killer and the age of the components.
You can with very careful selection substitute with higher spec components.

Could someone possibly suggest transistors with higher ratings for me?  or atleast tell me the parameters I need to match?

[tautech]

Quote from: tautech on Today at 09:25:35 AM
I would be tempted to replace all 4 transistors in the pre-amp stage.
If your tests are with removed devices, things are up the creek. 

Do you mean the final amp stage?  that is where Q407 - Q410 are from.  If that is not what you mean, can clarify please.

Correct, Q407-410 I miss-read the schematic

Quote from: tautech on July 03, 2014, 05:37:19 PM
Mostly voltage is the killer and the age of the components.
You can with very careful selection substitute with higher spec components.

Could someone possibly suggest transistors with higher ratings for me?  or atleast tell me the parameters I need to match?

Gain, Frequency, voltage, pin-out etc
As you have taken a while to work through this, what's a bit longer.
Take your time to hunt out replacements, X match them, Google substitutes etc

[onlooker]

I think a little LTspice exercise can help the diagnoses as to what should not be concentrated on. Yes, for the fun of it, I extracted a simplified version of the upper half of the output stage and did the simulation.

The simulation is done for a DC level sweeping.
 
The results are:

1). Bf = 5 to 200 for Q407 and Q408 has no significant impact on the performance of the output stage, (which has a voltage gain of ~6737). That is, they are non-critical. The much more critical things are the value drifts of the resistors in the feedback loop.

2). The proper Vin range is ~[2,5.5] V, for which Y+ linearly swings from about 130V to 10V. The 0V trace on screen should appear at Y+ ~= Y- ~=70V.

If I were doing the repair, I probably would not take out all the parts without checking these reference operating voltages first.

[ZaUcY]

First off, thanks for the simulation onlooker!

1). Bf = 5 to 200 for Q407 and Q408 has no significant impact on the performance of the output stage...

I am still confused on how you came to this conclusion, it looks like you hard coded the model to have a Bf = 5 for PNP and Bf = 20 for NPN.  Did you change the Bf value in the model and compare results?

The much more critical things are the value drifts of the resistors in the feedback loop.

Are you referring to the resistor values I highlighted yellow in my table, or are you just simply stating the correct resistor values are critical?  I still haven't pulled the resistors from the circuit to confirm they are bad yet.

2). The proper Vin range is ~[2,5.5] V

This is necessarily the proper Vin on my scope?

If I were doing the repair, I probably would not take out all the parts without checking these reference operating voltages first.

Unfortunately I have already pulled out the four transistors, but maybe that was good, because Q407 is more than likely a bad trans.  I have found new 2SA1360 and 2SC3423 trans on ebay that I have purchased to replace the transistors.  I had a look on digikey for upgrades, but I couldn't find a TO-126 package transistor with a high enough transition frequency.

I still plan on performing your DC levels check once I get all the components placed back in.

thanks again onlooker!

[onlooker]

Quote from: onlooker on Yesterday at 12:03:37 PM
   

1). Bf = 5 to 200 for Q407 and Q408 has no significant impact on the performance of the output stage...

I am still confused on how you came to this conclusion, it looks like you hard coded the model to have a Bf = 5 for PNP and Bf = 20 for NPN.  Did you change the Bf value in the model and compare results?

Bf = 5 for Q407 and Bf = 20 for Q408 were using the numbers you provided; Bf=200 was the datasheet number you or some other earlier poster quoted. I did tested the setup using Bf=5 and up for both transistors  (and also varied values of other components). The voltage gain remained about 37.

This is no surprise so long as the open loop gain is much larger than 37; Theoretically, for any Bf>0 (say 0.1 or 0.5) for Q408, the open loop gain can be infinite if the collector CC sink can be assumed ideal. Then, these are just DC analysis that may be helpful for get your 0V trace back if my quick simulation was correct.

Quote from: onlooker on Yesterday at 12:03:37 PM
   

The much more critical things are the value drifts of the resistors in the feedback loop.

Are you referring to the resistor values I highlighted yellow in my table, or are you just simply stating the correct resistor values are critical?  I still haven't pulled the resistors from the circuit to confirm they are bad yet.

The main feedback is through R415 in this part of schematics. Then, many other things can go wrong including the transistors you took out.

But, the 1st thing to do should be to determine if the problem is in this area by simply checking the voltage on Vin and Y+ on the schematics I showed (with auto trig on, no input on BNC and varying V pos,...). To be safe, you may clip the leads of your DMM in place with hands off.
 

Quote from: onlooker on Yesterday at 12:03:37 PM
   

2). The proper Vin range is ~[2,5.5] V

This is necessarily the proper Vin on my scope?

Vin is the test point marked on the schematics in my early post.  [2,5.5] V is the simulation findings for having a linear Y+. It is pretty stable when varying Bf and other things.

In any case, it can be served as a reference range for you, even if your scope has a range that did not fall exactly on it.

Quote from: onlooker on Yesterday at 12:03:37 PM
   

If I were doing the repair, I probably would not take out all the parts without checking these reference operating voltages first.

Unfortunately ...

Then, hope for the best. Be gentle with the pads and traces on "older" PCBs.
 

[ZaUcY]

Thanks for the further elaboration!  I just noticed the Vin on the graph

I will let you guys know what happens when I get the new transistors in solder everything back together, fortunately the solder pads have stayed intact thus far; only the solder mask is giving me trouble, it's fragile.

I will try all the DC measurements once it is all back together.

[ZaUcY]

For future reference, I have contacted BK Precision on the replacement transistors they use for Q407-Q410 and received the following response.

The A1360 have been replaced with NTE 374 and the C3423 with the NTE 373. Hope this helps.

Regards,
-Chris

[ZaUcY]

Hey guys!

Back with awesome news!  I received my Toshiba 2SA1360 and 2SC3423 pairs in the mail, from Germany, purchased off Ebay.  I soldered them into my scope, and soldered all the diodes and resistors back in that had one end de-soldered, and the scope is working now!!!!!

I have posted some pictures of using both channels, using component test on an LED and 10uF electrolytic, measuring a 10MHz 50mV p2p sine wave, and measuring the cal signal.  Everything seems to be working as expected (including everything that I did not photo) except the rising and falling edge of the cal signal looks funky.  I have never seen edges with variable intensities like the one shown in the picture below.  I don't have much experience with analog scopes, so if anyone else with more experience can elaborate, I would really appreciate it!

I haven't tried to measure signals or check calibration of the scope yet, unfortunately I'm writing this post on a train and only had time to solder the parts in this morning and take pictures.  So in about a week, I am going to be focusing on calibrating the scope.  Really excited it's working!!!  thanks for all the help guys!!  I now see that debugging a scope is a very typical and systematic approach, like everything else in EE, but I had no clue how scopes really work so it was daunting for me in the beginning.

Does anyone have tips for going about calibrating my scope?  I should have a reference scope, and function generator when trying to calibrate mine (college lab).  Unfortunately I have no clue on the calibration of these devices, and the reference scope will likely only be 100MHz.

thanks again

P.S.  I bought some extra new NTE transistors as well, in case these transistors die again

[tautech]

  
It is likely the scope is not far from accurate calibration.
Mainly I would check vertical accuracy.
There might be 3 adjustments for gain.
1 for each channel and another for overall gain.
A proper service manual will describe the procedure.

You can use a recently "in cal" scope for comparison and mostly this is quite good enough for hobby work.

Sometimes the Cal signal is adjustable(Amplitude & Freq) and it can be useful to get this exact for a quick reference in the future.

Mark any adjustment pots for a "return to" reference.

[ZaUcY]

Everything seems to be working as expected (including everything that I did not photo) except the rising and falling edge of the cal signal looks funky.  I have never seen edges with variable intensities like the one shown in the picture below.  I don't have much experience with analog scopes, so if anyone else with more experience can elaborate, I would really appreciate it!

I emailed BK Precision about this matter, and here was their response.

This is normal for analog oscilloscopes. The rise and fall is very fast, so it will exhibit those faint string of lines you see. The intensity of the light is based on the speed of the electron beam being traced on the phosphor screen. Adjusting the sweep speed setting will affect what you see. If it is slow, the vertical lines may not even be visible.

Makes sense!  I'm sure you guys already knew this.

[ZaUcY]

I took my scope into an old college lab, and everything seemed pretty calibrated.  I only really knew what about one third of the pots did, but all of those pots seemed to be calibrated pretty good (according to reference scope and function generator).  The only thing I didn't like is that everything agreed in lower frequencies, but when I ran a 10Mhz signal, amplitudes did not agree.  It maybe had something to do with the front end capacitors, but I didn't really have access to them  

Scope works, it seems calibrated, I am a happy engineer