1. Home
  2. Electronics Design
  3. Need some Circuit analysis/troubleshooting help

Need some Circuit analysis/troubleshooting help

Hi... I'm struggling with a repair issue and could use some help. I've placed some schematics/pics at this link: picture Link:

formatting link

The problem that I am having trouble with is a bipolar amp/bandpass filter used on a log amp board from an old HP network analyzer. This analyzer uses two boards, which are identical, for the reference channel and the A,B channels (muxed). One board functions without issue, the other has a random noise problem.

If you look at the schematic in the link, there is a test point, TP 2, which represents the location of the noise busts shown on the schematic. These noise bursts are random in nature and I have verified that they are not dur to mechanical vibration. The problematic circuit card is being test/analyzed on the bench using a bench +/-15V supply. The schematic shows the bias point at this TP2 to be -.5V and the 'good' board has been verified to be true.

My analysis shows that this bias point should be about -0.9Vdc, not -0.5Vdc. On initial application of the power supplies, the bias point starts off at about -.8V on the good board and then over several minutes settles to -.5V and all is happy. On the bad board the voltage starts off about -.7V but settles to around -.2V or so. The bad board also starts exhibiting these noise bursts which are causing me problems on the low end of the dynamic range of the analyzer.

Initially, I thought that I was having a leaky npn and or pnp in this circuit. The schematic also indicates in red areas where I have removed various parts in an effort to completely isolate this area from the remaining circuitry. I've also 'turned off' many of the other stages (lifting bias resistors, etc.) in an effort to make sure that this stage is the one really at fault. I also shorted out the 100KHz inductor (250uH) to make sure that it wasn't receiving noise bursts from some external means.

The normal troubleshooting has been done: All electrolytics have been tested for ESR/Cap Resistance measurements have been verified. Extra power supply decoupling has been done to insure clean power. The HP transistors have been cross referenced and replaced, thinking that there was some leakage issues, etc. btw.. these are 2N3904/06 type parts.

Ok... That is best that I can describe my problem. At this point what I need is to understand WHY the bias point is not where I believe it should be and why does it take several minutes to stabilize? Heat doesn't seem to be an issue here. If there is an instability in this feedback amp, I could use some pointers as to how an correct it.

thanks for reading my gibberish and look forward to hearing back from someone. Jim

Reply to
Jim Flanagan
Loading thread data ...

At first glance, R59 + R54 constitute positive feedback!

Under what specification are you to set R54?

Might be a cantankerous adjustment. ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |

      Remember: Once you go over the hill, you pick up speed
Reply to
Jim Thompson

thanks for the response.

The circuit being analyzed is a 1KHz/10KHz simple bandpass. The series resistor (R52) is being switched in/out. The gain of this bp/amp is only about 2 and R54 is used to compensate for the difference in gain when switching between the 1 & 10 KHz bandwidths.

jim

Reply to
Jim Flanagan

You have a component shifting and and I bet it's a bad R or even a cap that is changing value in operation.. Having a cap pass a ESR check does not mean its good. It could have a voltage break down problem with in it's spec's, you would need to apply rated voltage across the cap for this and have a current meter in line.

Most ESR meters use a very low voltage, so you won't see this.

What does work a lot how ever, is component freeze spray on R's and C's

have your meter on the TP when doing so. For noise like that, I would first try the resistors. what for the noise and spray!

Jamie

Reply to
Jamie

I'm just pointing out that the setting of R54 affects bias. Or maybe some other element in that area is defective. Compare bias on all nodes between good and bad boards. ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |

      Remember: Once you go over the hill, you pick up speed
Reply to
Jim Thompson

Since the base of the NPN is at DC ground (through the 250 uH) I figured R54 was only going to affect the AC gain and not the DC bias. Am I mistaken?

Jim

Reply to
Jim Flanagan

Jim-

After all you've been through, I'd look for a plated-through hole that isn't.

Fred

Reply to
Fred McKenzie

Oooops! You are correct for DC. But suppose it is popping in and out of oscillation... that would affect apparent bias. You did mention observing some spurious behavior. ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |

      Remember: Once you go over the hill, you pick up speed
Reply to
Jim Thompson

e

=A0 =A0 ...Jim Thompson

You're correct, R54 does not affect DC bias.

You said you shorted L1. a) Did that eliminate the noise bursts at TP2? b) What do these 'noise' bursts look like (amplitude, polarity, frequency (if applicable)) etc.? Looking at that should tell you whether the circuit's oscillating or something else.

You need more clues. If the 'noise' burst includes a d.c. shift, the polarity of the d.c. shift might tell you something, etc.

Sounds mechanical to me, like an intermittent connection. Jamie's freeze-spray idea's good for finding those.

-- Cheers, James Arthur

Reply to
dagmargoodboat

Yes.. the spurious behavior is the main problem, for sure. I've actually removed the fixed series resistor in the feedback path and saw no difference in both bias and spurious behavior. I'm puzzled to say the least. Much work to do, it appears. When I compare all voltages around the circuit, they are all correct (except the TP2). During times like this, I take a day or so off to *think*, then I dive back in.

In your experience, does the circuit topology look odd to you? Have you seen similar circuits like this before? I appreciate the banter that you are having with me on this. Soothes the soul when frustration sets in... Jim

Reply to
Jim Flanagan

Hi, Jim -

Rock the wiper of R54 a little while looking at TP2.

I have found lots of noisy pots in older equipment. They are usually most noisy where the wiper has rested on the element for so many years.

Anyway, good luck.

Cheers, John

Reply to
John - KD5YI

Great! After a day of building shelving, we just returned from la Ragazze, antipasto and wine :-) Will resume tomorrow. ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |

      Remember: Once you go over the hill, you pick up speed
Reply to
Jim Thompson

If you have erratic operation of a Hi Z DC circuit, one thing that is suspect, is contamination of the PWB. Clean it with rubbing alcohol and let it fully dry before applying power. '

let us know what you find... Mark

Reply to
Mark

be

That was my suggestion, maybe measure it's value and replace with a resistor.

George H.

Reply to
George Herold

e

Hi Jim

Been there - done that ;-)

-

Temorarily remove the multi stage feedback:

  • remove R60

  • remove R59

  • connect R54 to a resistor to +15V; adjust value, so TP2 is approximatly -0,5V (The value will drift and fluctuate).

Now you can check stage induce noise.

-

In my search I found an old silicium transistor that made the noise- bursts, but it might as well be one of the other components.

When you have found and replaced the offending component, you restore the circuit feedback.

Reply to
Glenn

Also disconnect R50, R52 while checking the two individual stages.

Reply to
Glenn

I have also seen many crackling soldering points, where the bad/ degraded point had a crater rim around the terminal when looking through a magnifying glass (x10).

The bad/degraded soldering points was typically carrying high currents and/or high temperature. Tube sockets, power resistors, power diodes and power transistors.

I also have a GF759(almost the same af BF459) where the Collector disconnected spontaniously inside, but the base and emitter was fine.

Reply to
Glenn

Have you manipulated R54, or subbed it out? Perhaps a noisy wiper.

When you are measuring the -0.5V bias test point, is this with/without signal? ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |

      Remember: Once you go over the hill, you pick up speed
Reply to
Jim Thompson

I can sort of match your symptoms with a leakage path to ground from the base of T2.

Crud on the board? ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |

      Remember: Once you go over the hill, you pick up speed
Reply to
Jim Thompson

Well, I've noticed that the bias is extremely sensitive to the +/-15 supplies. These supplies are not regulated at the board. There are series inductors where the power enters that has about 7 ohms of resistance forcing a DC drop that is dependent on all the rest of the circuitry. So in reality the voltage at the problem stage is around 14.5Vdc. If you play slightly with these supply voltages you can tweak the bias to about -.5V. I ended up adding some extra bypassing (.1uf) around the circuit of interest with no improvement.

Btw.. the R54/R59 network has been removed with the same problems.

What I have done today is to lift the legs of the resistors where the

+/-15 supplies are fed and directly connected the bench supplies to these lifted legs. This leaves ALL of the rest of the board unpowered. Same random noise issues. Weird ...

To answer your question, the measurements are taken with no 100 KHz signal applied to the board.

It seems HIGHLY unlikely, but could any of the resistors in the network have any known condition that would cause this type of noise? They all measure OK with the ohmmeter.

It may be time to punt.. again... appreciate your help. Jim

Reply to
Jim Flanagan

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.