开云体育

My faithful 3585a has developed a problem. When a BNC cable is connected to either the 50/75 ohm or 1 megohm input and the cable shield is connected to ground contact of another instrument or the ground contact from a wall outlet, spikes walk across the screen. These spikes are about 20dBm above the noise floor.

If I set the instrument to its default settings, I see the spikes walk across the screen. If I save the waveform to B (where it no longer walks), I count 24 spikes. On many start-stop frequency settings I see the same count of spikes. However, on some I see 48 spikes; e.g. a start frequency of 1MHz and a stop of 2MHz.

Anyone have a suggestion of how to diagnose and repair this problem?

Can the work be done without extender boards?

Thanks,

Gordon, KA2NLM

If you can get hold of one of those ancient 2 to 3 prong AC adapters it is
good for breaking ground loops. If you have one try it and see if the
spikes go away.

Regards.

Max. K 4 O DS.

Email: max@...

Gordon wrote.

My faithful 3585a has developed a problem. When a BNC cable is connected
to either the 50/75 ohm

or 1 megohm input and the cable shield is connected to ground contact of
another instrument or the ground contact from a wall outlet, spikes walk
across the screen. These spikes are about 20dBm above the noise floor.

Maybe your home has been equipped with a smart electric meter. Mine has and
it transmits on about 160 kHz and has strong harmonics up to about 10 MHz.
It's a very wide band digital signal that is on the power lines. It can
easily be induced into any ground loop which is what you are creating when
you connect two pieces of grounded equipment together with a coaxial cable.
If you can get hold of one of those ancient 2 to 3 prong AC adapters it is
good for breaking ground loops. If you have one try it and see if the
spikes go away.

Regards.

Max. K 4 O DS.

Email: max@...

Transistor site
Vacuum tube site:
Woodworking site

Music site:

To subscribe to the fun with transistors group send an email to.
funwithtransistors-subscribe@...

To subscribe to the fun with tubes group send an email to,
funwithtubes-subscribe@...

To subscribe to the fun with wood group send a blank email to
funwithwood-subscribe@...

----- Original Message -----
From: "gldinnc"
To:
Sent: Wednesday, January 09, 2013 4:24 PM
Subject: [hp_agilent_equipment] Failing 3585A Spectrum Analyzer

My faithful 3585a has developed a problem. When a BNC cable is connected
to either the 50/75 ohm

or 1 megohm input and the cable shield is connected to ground contact of
another instrument or the ground contact from a wall outlet, spikes walk
across the screen. These spikes are about 20dBm above the noise floor.

If I set the instrument to its default settings, I see the spikes walk
across the screen. If I save the waveform to B (where it no longer walks),
I count 24 spikes. On many start-stop frequency settings I see the same
count of spikes. However, on some I see 48 spikes; e.g. a start frequency
of 1MHz and a stop of 2MHz.

Anyone have a suggestion of how to diagnose and repair this problem?

Can the work be done without extender boards?

Thanks,

Gordon, KA2NLM




------------------------------------

Yahoo! Groups Links

My faithful 3585a has developed a problem. When a BNC cable is connected to either the 50/75 ohm or 1 megohm input and the cable shield is connected to ground contact of another instrument or the ground contact from a wall outlet, spikes walk across the screen. These spikes are about 20dBm above the noise floor.

If I set the instrument to its default settings, I see the spikes walk across the screen. If I save the waveform to B (where it no longer walks), I count 24 spikes. On many start-stop frequency settings I see the same count of spikes. However, on some I see 48 spikes; e.g. a start frequency of 1MHz and a stop of 2MHz.

Anyone have a suggestion of how to diagnose and repair this problem?

Can the work be done without extender boards?

Thanks,

Gordon, KA2NLM

In response to the replies to date:

1. The spikes stop moving if I trigger off the line.
2. The number of spikes is not depended on frequency; therefore the frequency between spikes varies.
3. The 3585a worked fine until recently. No wiring changes have been made. Two different runs back to the breaker box have been used with the same result.

More suggestions please,

Gordon, KA2NLM

--- In hp_agilent_equipment@..., "gldinnc" wrote:

My faithful 3585a has developed a problem. When a BNC cable is connected to either the 50/75 ohm or 1 megohm input and the cable shield is connected to ground contact of another instrument or the ground contact from a wall outlet, spikes walk across the screen. These spikes are about 20dBm above the noise floor.

If I set the instrument to its default settings, I see the spikes walk across the screen. If I save the waveform to B (where it no longer walks), I count 24 spikes. On many start-stop frequency settings I see the same count of spikes. However, on some I see 48 spikes; e.g. a start frequency of 1MHz and a stop of 2MHz.

Anyone have a suggestion of how to diagnose and repair this problem?

Can the work be done without extender boards?

Thanks,

Gordon, KA2NLM

By the way, I have scanned the power supply trouble shooting section and it seems like most of the likely steps call for the use of extender boards. Is it realistic to attempt repair without extender boards?

Gordon, KA2NLM

--- In hp_agilent_equipment@..., "gldinnc" wrote:

In response to the replies to date:

1. The spikes stop moving if I trigger off the line.
2. The number of spikes is not depended on frequency; therefore the frequency between spikes varies.
3. The 3585a worked fine until recently. No wiring changes have been made. Two different runs back to the breaker box have been used with the same result.

More suggestions please,

Gordon, KA2NLM

Hi Gordon,

Since you managed to stop the spikes from moving with LINE TRIGGER, I would say that you have a power supply filtering problem. I don't remember the inside of the 3585 any more, but I suspect that you could easily use a scope to check all the power supplies for ripple. That would be the first thing I would do and you may not need the extenders for that.

Vladan

--- In hp_agilent_equipment@..., "gldinnc" wrote:

By the way, I have scanned the power supply trouble shooting section and it seems like most of the likely steps call for the use of extender boards. Is it realistic to attempt repair without extender boards?

Gordon, KA2NLM

--- In hp_agilent_equipment@..., "gldinnc" wrote:

In response to the replies to date:

1. The spikes stop moving if I trigger off the line.
2. The number of spikes is not depended on frequency; therefore the frequency between spikes varies.
3. The 3585a worked fine until recently. No wiring changes have been made. Two different runs back to the breaker box have been used with the same result.

More suggestions please,

Gordon, KA2NLM

Thanks to Vladan, Ken and Ed for their suggestions.

I did a little more testing. Here is what I found:

1. With a battery operated source, everything seems fine.
2. If an external ground is connected to the ground of the tracking generator BNC or the external trigger BNC, no spikes appear.
3. Connecting external ground to the 1meg or 50/75 ground generates spikes.
4. If the span is less than 500kHz the spikes disappear (or seem minor), but if the span goes from 499,999 to 500,000 they hop up to the reported level.

While I can pull the covers off and begin to test power supply ripple, I would hope to get a little more directed.

Thoughts?

Gordon, KA2NLM

Hi Gordon,

I assume the battery operated source is a battery with an inverter that creates 120Vac/60Hz, right? Does the inverter output have a safety ground that is connected to earth?

If not, please do the following experiment. Disconnect everything from the 3585 except for the power cord and plug it into your usual ac line socket. Turn the 3585 on. Take a multimeter and put it on AC Volts . Attach one lead of the meter to something that you know is grounded (earth) and touch the other meter lead to the BNC shells and any conductive chassis parts on the 3585. Is there any measurable voltage?

Vladan

--- In hp_agilent_equipment@..., "gldinnc" wrote:

Thanks to Vladan, Ken and Ed for their suggestions.

I did a little more testing. Here is what I found:

1. With a battery operated source, everything seems fine.
2. If an external ground is connected to the ground of the tracking generator BNC or the external trigger BNC, no spikes appear.
3. Connecting external ground to the 1meg or 50/75 ground generates spikes.
4. If the span is less than 500kHz the spikes disappear (or seem minor), but if the span goes from 499,999 to 500,000 they hop up to the reported level.

While I can pull the covers off and begin to test power supply ripple, I would hope to get a little more directed.

Thoughts?

Gordon, KA2NLM

Vladan,

I was a little too brief (confusing) in my first point. I was using a battery operated signal source to put a test signal on the 3585a, not to power the 3585a.

I have conducted the suggested test using a Fluke 189. I get no voltage across any of the BNC shells or the 3585a chassis.

Thanks for sticking with me on this. I really would like to get it to work.

Gordon.


--- In hp_agilent_equipment@..., "pianovt" wrote:

Hi Gordon,

I assume the battery operated source is a battery with an inverter that creates 120Vac/60Hz, right? Does the inverter output have a safety ground that is connected to earth?

If not, please do the following experiment. Disconnect everything from the 3585 except for the power cord and plug it into your usual ac line socket. Turn the 3585 on. Take a multimeter and put it on AC Volts . Attach one lead of the meter to something that you know is grounded (earth) and touch the other meter lead to the BNC shells and any conductive chassis parts on the 3585. Is there any measurable voltage?

Vladan

--- In hp_agilent_equipment@..., "gldinnc" wrote:

Thanks to Vladan, Ken and Ed for their suggestions.

I did a little more testing. Here is what I found:

1. With a battery operated source, everything seems fine.
2. If an external ground is connected to the ground of the tracking generator BNC or the external trigger BNC, no spikes appear.
3. Connecting external ground to the 1meg or 50/75 ground generates spikes.
4. If the span is less than 500kHz the spikes disappear (or seem minor), but if the span goes from 499,999 to 500,000 they hop up to the reported level.

While I can pull the covers off and begin to test power supply ripple, I would hope to get a little more directed.

Thoughts?

Gordon, KA2NLM

Vladan,

I was a little too brief (confusing) in my first point. I was using a battery operated signal source to put a test signal on the 3585a, not to power the 3585a.

I have conducted the suggested test using a Fluke 189. I get no voltage across any of the BNC shells or the 3585a chassis.

Thanks for sticking with me on this. I really would like to get it to work.

Gordon.

Hi Gordon,

If I were troubleshooting this problem, I would proceed as follows:

1. Open the top cover and check the power supplies for ripple. I would do this first, because it's the easiest thing to do and a bad filter capacitor might be the problem. On page 8-10 of Service Manual Vol. 2, you will see all the test points where you can probe with the scope and measure D.C.

2. If that all looks fine (no ripple on the supplies), my next guess would be to look at Figure 10-38 (service manual Vol. 2) "Power Distribution". In the upper right corner is what looks like a common mode filter T2. This filter is hooked up to the U1 and U3 regulators. Check the output of the regulators for ripple. Look at T2 to see if something is wrong with it. The reason I zoomed in on that circuit is because terminals 1 and 4 of T2 isolate the power supply ground from the input board ground. On the schematics, the power supply ground is called GND, and the ground for the front end is called GND1. In other words, T2 is where the BNC shells of your input connectors get their ground (GND1) and that ground is not the same as the power supply ground (GND). Using an Ohm meter, you should however see continuity (probably 0 Ohms) between the shells of the BNCs (GND1) and the power supply ground (GND).

Other than that, I don't know what to suggest. Best luck,

Vladan

--- In hp_agilent_equipment@..., "gldinnc" wrote:

Vladan,

I was a little too brief (confusing) in my first point. I was using a battery operated signal source to put a test signal on the 3585a, not to power the 3585a.

I have conducted the suggested test using a Fluke 189. I get no voltage across any of the BNC shells or the 3585a chassis.

Thanks for sticking with me on this. I really would like to get it to work.

Gordon.

Hi Gordon,

If I were troubleshooting this problem, I would proceed as follows:

1. Open the top cover and check the power supplies for ripple. I would do this first, because it's the easiest thing to do and a bad filter capacitor might be the problem. On page 8-10 of Service Manual Vol. 2, you will see all the test points where you can probe with the scope and measure D.C.

2. If that all looks fine (no ripple on the supplies), my next guess would be to look at Figure 10-38 (service manual Vol. 2) "Power Distribution". In the upper right corner is what looks like a common mode filter T2. This filter is hooked up to the U1 and U3 regulators. Check the output of the regulators for ripple. Look at T2 to see if something is wrong with it. The reason I zoomed in on that circuit is because terminals 1 and 4 of T2 isolate the power supply ground from the input board ground. On the schematics, the power supply ground is called GND, and the ground for the front end is called GND1. In other words, T2 is where the BNC shells of your input connectors get their ground (GND1) and that ground is not the same as the power supply ground (GND). Using an Ohm meter, you should however see continuity (probably 0 Ohms) between the shells of the BNCs (GND1) and the power supply ground (GND).

Other than that, I don't know what to suggest. Best luck,

Vladan

--- In hp_agilent_equipment@..., "gldinnc" wrote:

Vladan,

I was a little too brief (confusing) in my first point. I was using a battery operated signal source to put a test signal on the 3585a, not to power the 3585a.

I have conducted the suggested test using a Fluke 189. I get no voltage across any of the BNC shells or the 3585a chassis.

Thanks for sticking with me on this. I really would like to get it to work.

Gordon.

What I find curious, is that the spikes on the display only happen when the frequency span is 500,000 Hz or higher. Thoughts on this?

Vladan,

Here are the results of my testing :(

1. The test points on A71-A74 all are well within tolerance and show no ripple. There is some pulse noise in the 20KHz range, but I would guess that is normal (I have a 200 Meg Scope).

2. I checked the power out of A76 and there is no ripple on J3, J7, J9 and J10. Again there is a little high frequency pulse noise.

3. Resistance is nominally 0 between A76 J10 and the BNC shell.

What I find curious, is that the spikes on the display only happen when the frequency span is 500,000 Hz or higher. Thoughts on this?

Thanks,

Gordon.

--- In hp_agilent_equipment@..., "pianovt" wrote:

Hi Gordon,

If I were troubleshooting this problem, I would proceed as follows:

1. Open the top cover and check the power supplies for ripple. I would do this first, because it's the easiest thing to do and a bad filter capacitor might be the problem. On page 8-10 of Service Manual Vol. 2, you will see all the test points where you can probe with the scope and measure D.C.

2. If that all looks fine (no ripple on the supplies), my next guess would be to look at Figure 10-38 (service manual Vol. 2) "Power Distribution". In the upper right corner is what looks like a common mode filter T2. This filter is hooked up to the U1 and U3 regulators. Check the output of the regulators for ripple. Look at T2 to see if something is wrong with it. The reason I zoomed in on that circuit is because terminals 1 and 4 of T2 isolate the power supply ground from the input board ground. On the schematics, the power supply ground is called GND, and the ground for the front end is called GND1. In other words, T2 is where the BNC shells of your input connectors get their ground (GND1) and that ground is not the same as the power supply ground (GND). Using an Ohm meter, you should however see continuity (probably 0 Ohms) between the shells of the BNCs (GND1) and the power supply ground (GND).

Other than that, I don't know what to suggest. Best luck,

Vladan

--- In hp_agilent_equipment@..., "gldinnc" wrote:

Vladan,

I was a little too brief (confusing) in my first point. I was using a battery operated signal source to put a test signal on the 3585a, not to power the 3585a.

I have conducted the suggested test using a Fluke 189. I get no voltage across any of the BNC shells or the 3585a chassis.

Thanks for sticking with me on this. I really would like to get it to work.

Gordon.

Hi Gordon,

This is almost certainly related to the way the LO is structured. Generally speaking, the LO has to make some changes as the span increases. In your case, for spans that are 500 kHz and larger, the LO uses a different method of generating the signal. I will look at the manual, but this may not be the easiest way to find a bad part, especially since you see the problem (to a lesser extent) in narrower spans as well.

Vladan

--- In hp_agilent_equipment@..., "gldinnc" wrote:

What I find curious, is that the spikes on the display only happen when the frequency span is 500,000 Hz or higher. Thoughts on this?