E4407B Tracking Generator Repair

[smgvbest]

I installed a tracking generator I got off ebay into my repaired E4407B and after making the LO and 600Mhz Reference Cables fire it up and found it does have an issue.
When it fire up and enabled the TG I received a Source TG Unlevel,   i double checked my cabled and snugged up the LO cable at the LOIS side and the Source TG Unlevel went away and was replaced with a Source Unlevel.   Doing some checking at 0.5DdB intervals I checked from 4dB to -66dB and got the table below.

Interesting that at a fairly regular interval it goes from Source Unlevel to No Error (Pass in the table)

Code: [Select]

Start End         Range Status
3.0         -6.5         -9.5         Source Unlevel
-7.0         -9.5         -2.5         Pass
-10.0 -14.5 -4.5         Source Unlevel
-15.0 -17.5 -2.5         Pass
-18.0 -22.0 -4.0         Source Unlevel
-22.5 -25.5 -3.0         Pass
-26.0 -29.5 -3.5         Source Unlevel
-30.0 -33.5 -3.5         Pass
-34.0 -37.5 -3.5         Source Unlevel
-38.0 -41.5 -3.5         Pass
-42.0 -45.5 -3.5         Source Unlevel
-46.0 -49.5 -3.5         Pass
-50.0 -53.5 -3.5         Source Unlevel
-54.0 -57.5 -3.5         Pass
-58.0 -61.0 -3.0         Source Unlevel
-61.5 -66.0 -4.5         Pass

At the suggestion of Jürgen over on the HP-Agilent-Keysight group.io group I disconnected the RF out to the attenuator and ran same test.   I got same result
the idea being is the issue with the LO on the control board or in the BITG which is not documented.

So I'm going to basics and starting with a visual inspection then checking the LO Path on the control board.

So what follows is a boat load of images of the control board.   after those images I will continue posting on testing

[smgvbest]

This is the top side of the control board E4403-60004

[smgvbest]

rest of top board

[smgvbest]

Board side of Control Board

[smgvbest]

Close up of the LO Amps and Pin Modulator sections

[smgvbest]

here's the Attenuator/cables and exterior of the BITG

[smgvbest]

so after getting all those images (yes from the phone not a real camera)
I checked all the FETS and PIN Diodes using the diode test feature of the DMM

attached at the relavant reading on the schematic sheets

All the FETS should be a ATF-13484,  the marking though are a mix of F25 and 101 on them

it looks like Q12, Q13,Q11,Q10 and Q9 may have been replaced which are all labled F25.
the one that stood out to me is Q15 though.
the other ones marked 101  has reading of 0.465/1.329vdc where as Q15 read .549/2.497vdc.   the question is, is Q15 the odd ball or is it Q14/Q16

That's where I am at this point

Any one with suggestions or feedback I welcome the input

[smgvbest]

Continuing the teardown aspect of the Tracking Generator,  next is the A2A2 BITG module.   This is the black box.  it's not documented in the CLIP

[smgvbest]

These pics are more for inspection of certain points on same board

[analogRF]

maybe I missed something but why dont you just simply start by checking the levels at the major LO connections? I mean the LO in/out levels at the control board and the LO in at the TG module (which is the same as LO out from control board)? just vary the center frequency at zero span
and vary the TG amplitude and monitor the LO levels at these 2 points (one at a time) it should be fairly easy to rule out one of the modules

since the error occurs at some intervals , it seems to me the problem could be either an electronic attenuator somewhere OR an electronic RF switch which changes the signal path depending on the level requested (more likely in my opinion). I think I saw a couple of them in one of the pictures. I need to look again. But the circuits are very simple and easy to diagnose. there is no need for schematics

[smgvbest]

Here's the PICs of control board for the BITG

[smgvbest]

and last two

[analogRF]

there doesn't seem to be any RF switches or electronic attenuators at first glance. I may have missed something though...but clearly you have several Minicircuits MCL29 gain blocks and HMC311 Hittite/analog device/ gain blocks and I would suspect them. You can start by checking their supply voltages based on their datasheets. if you can make an extension ribbon cable and bring the board out then it would be fairly easy to probe even with a home made 500Ohm RF probe and a spectrum analyzer or power meter at least.

[smgvbest]

maybe I missed something but why dont you just simply start by checking the levels at the major LO connections? I mean the LO in/out levels at the control board and the LO in at the TG module (which is the same as LO out from control board)? just vary the center frequency at zero span
and vary the TG amplitude and monitor the LO levels at these 2 points (one at a time) it should be fairly easy to rule out one of the modules

since the error occurs at some intervals , it seems to me the problem could be either an electronic attenuator somewhere OR an electronic RF switch which changes the signal path depending on the level requested (more likely in my opinion). I think I saw a couple of them in one of the pictures. I need to look again. But the circuits are very simple and easy to diagnose. there is no need for schematics

Hi,

So a few thing on that.   the only SA I have that can look at the LO is the E4407B.  when you enable the TG it turns on the LO to the TG and your range is reduced to 0-3Ghz.  if you go outside the 3Ghz range which you need to for the LO (starts around 3.9 Ghz if I recall) the TG LO is turned out so you can't view it on the SA itself.

Second item is there are 2 errors associated with the TG
Source LO Unlevel (whcih I am not getting) and points to the LO being a problem on the Control Board.
Source Unlevel (which I am getting) points to a problem on the A2A2 (BITG) board.   

I've disconnected and terminate the LO coming from the control board and still get same errors. 

The only reason to check those FETS and PIN diodes was because at least a few of the FETS had been replaced and that's me doing my due diligence it looking at them

the good thing with the BITG is it sits high enough that if I leave the shield off I can likely probe the board.

I've spent most my time right now documenting these board as there's nothing out there on the TG.

Now that I have its time for troubleshooting.

I welcome any advice I can get here.   RF is not my specialty by any means.   I think I see the signalpath (not trying to steal their name or anything) so there's not many places I can see it could go bad on the BITG

[analogRF]

some thoughts...

[analogRF]

maybe I missed something but why dont you just simply start by checking the levels at the major LO connections? I mean the LO in/out levels at the control board and the LO in at the TG module (which is the same as LO out from control board)? just vary the center frequency at zero span
and vary the TG amplitude and monitor the LO levels at these 2 points (one at a time) it should be fairly easy to rule out one of the modules

since the error occurs at some intervals , it seems to me the problem could be either an electronic attenuator somewhere OR an electronic RF switch which changes the signal path depending on the level requested (more likely in my opinion). I think I saw a couple of them in one of the pictures. I need to look again. But the circuits are very simple and easy to diagnose. there is no need for schematics

Hi,

So a few thing on that.   the only SA I have that can look at the LO is the E4407B.  when you enable the TG it turns on the LO to the TG and your range is reduced to 0-3Ghz.  if you go outside the 3Ghz range which you need to for the LO (starts around 3.9 Ghz if I recall) the TG LO is turned out so you can't view it on the SA itself.

Second item is there are 2 errors associated with the TG
Source LO Unlevel (whcih I am not getting) and points to the LO being a problem on the Control Board.
Source Unlevel (which I am getting) points to a problem on the A2A2 (BITG) board.   

I've disconnected and terminate the LO coming from the control board and still get same errors. 

The only reason to check those FETS and PIN diodes was because at least a few of the FETS had been replaced and that's me doing my due diligence it looking at them

the good thing with the BITG is it sits high enough that if I leave the shield off I can likely probe the board.

I've spent most my time right now documenting these board as there's nothing out there on the TG.

Now that I have its time for troubleshooting.

I welcome any advice I can get here.   RF is not my specialty by any means.   I think I see the signalpath (not trying to steal their name or anything) so there's not many places I can see it could go bad on the BITG

so clearly the problem is only inside the A2A2 and in that module the entire section behind the RF output (marked by yellow color below) is 0-3GHz and you should be able to probe it. The level problem is more than likely happening in this section and not in other sections (there are scenarios that it can be traced to other sections but the most likely cause of unlevel issue must be the section I have highlighted with yellow color here)

there are only two RF gain blocks (check their supply first) and if they are ok the only problem is the IGG hybrid which does the level detection
and RPP    if someone applied RF power or DC to the tracking gen, the IGG5 is gone 

just to make sure, you can also check the AD620 opamp that is most likely connected to the IGG5 and completes the level detection and probably generates the error signal based on some reference value.

[suj]

Maybe try to set the TG frequency quite low and try to check the level with an oscilloscope? In this case, I prefer the 3 GHz HP 85024A probe, but I have other SA's at my disposal. Maybe diode detector with additional resistor at the input??

[smgvbest]

so clearly the problem is only inside the A2A2 and in that module the entire section behind the RF output (marked by yellow color below) is 0-3GHz and you should be able to probe it. The level problem is more than likely happening in this section and not in other sections (there are scenarios that it can be traced to other sections but the most likely cause of unlevel issue must be the section I have highlighted with yellow color here)

there are only two RF gain blocks (check their supply first) and if they are ok the only problem is the IGG hybrid which does the level detection
and RPP    if someone applied RF power or DC to the tracking gen, the IGG5 is gone 

just to make sure, you can also check the AD620 opamp that is most likely connected to the IGG5 and completes the level detection and probably generates the error signal based on some reference value.

Thank you for the analysis. that's very helpfull. 
I've been collecting datasheets so here they are for reference


my thought was with HMC346, the variable attenuator.
The area you mention is much easier to probe though.  my other sa's go to 1.5Ghz (Rigol) and 2.7 Ghz(Agilent Base Station)

[analogRF]

I would start by checking the output section. any unlevel issue most likely happens in there (+ probably the opamp on the upper right of the IGG5)

if that section is fine, then we can think of other scenarios.

1.5GHz SA is fine. I believe your unlevel errors happen at any frequency, right? or do they happen at certain frequencies?!

if they happen at certain frequency ranges, then perhaps that HMC346 is a better place to start. But I thought your errors were happening at all frequencies

[analogRF]

not sure if the golden mixer is HMC129. The packaging does not match. still it could be an old version of hmc129 with different specs...

another peculiarity here is that MNA6 (bottom right) is specified as 0.5-2.5GHz amplifier. So the signal in that path cannot be 3.92GHz as we expect.
I am thinking maybe that path is 1.96GHz and the golden mixer H129, actually does a harmonic mixing (mixing with the second harmonic of 1.96) ?
I wish you could do some tests and verify what frequency is there..

[analogRF]

not sure if the golden mixer is HMC129. The packaging does not match. still it could be an old version of hmc129 with different specs...

another peculiarity here is that MNA6 (bottom right) is specified as 0.5-2.5GHz amplifier. So the signal in that path cannot be 3.92GHz as we expect.
I am thinking maybe that path is 1.96GHz and the golden mixer H129, actually does a harmonic mixing (mixing with the second harmonic of 1.96) ?
I wish you could do some tests and verify what frequency is there..

or maybe the signal is 1.96GHz at the MNA6 amplifier but then they multiply it by 2 using something similar to a diode at the bottom edge of the PCB right before the bandpass filter. So after that the BPF, filters in the 3.92GHz. This looks more plausible....that BPF must be very sharp with such widely spaced resonators.

[smgvbest]

not sure if the golden mixer is HMC129. The packaging does not match. still it could be an old version of hmc129 with different specs...

another peculiarity here is that MNA6 (bottom right) is specified as 0.5-2.5GHz amplifier. So the signal in that path cannot be 3.92GHz as we expect.
I am thinking maybe that path is 1.96GHz and the golden mixer H129, actually does a harmonic mixing (mixing with the second harmonic of 1.96) ?
I wish you could do some tests and verify what frequency is there..

What specifically would help the most to test?   I will work on doing that.
It is a 600Mhz in to this unit in the upper right.
the section you mark oscill? freq?   there are 3 pins this coming from the other side of the board as seen in the attached image

the other image is of the signals on the 10pin header input coming from the control board.  ALC_MON goes back to a OPAMP on the control board
ALC_EXT is suppose to be on same page but I cant' locate it.

[analogRF]

not sure if the golden mixer is HMC129. The packaging does not match. still it could be an old version of hmc129 with different specs...

another peculiarity here is that MNA6 (bottom right) is specified as 0.5-2.5GHz amplifier. So the signal in that path cannot be 3.92GHz as we expect.
I am thinking maybe that path is 1.96GHz and the golden mixer H129, actually does a harmonic mixing (mixing with the second harmonic of 1.96) ?
I wish you could do some tests and verify what frequency is there..

What specifically would help the most to test?   I will work on doing that.
It is a 600Mhz in to this unit in the upper right.
the section you mark oscill? freq?   there are 3 pins this coming from the other side of the board as seen in the attached image

the other image is of the signals on the 10pin header input coming from the control board.  ALC_MON goes back to a OPAMP on the control board
ALC_EXT is suppose to be on same page but I cant' locate it.

what pins are coming out of that area? can you mark them? do they connect to the other board?

w.r.t. tests, as I mentioned, i would start by just looking at the yellow marked area with an RF probe (even a home made 450ohm series resistor with RG316 cable. you can use two back to back female sma-to-board connectors and solder the SMD resistor between the center pins of the two and solder the surrounding pins together to make a cage around the resistor and then insert a pin inside the center and solder a solid wire to the body as ground pin. The other side is connected to your measuring device with standard RG316 SMA cable. I'll post a picture of mine tomorrow. It has worked for me flawlessly up to 7GHz before)
or if you can even check at low frequencies by an oscilloscope. Check to see if you have a good  signal at the input to the IGG5 hybrid or not when the error occurs. and if not, work your way back to the two amplifiers. Check their supply voltage (seems to me it should be 12V by the value of series resistor they used for biasing)
one thing that is important is to see if your unlevel errors happen at all frequencies or only at certain ranges. because that makes a big difference in strategy

[Jay_Diddy_B]

Hi,

I don't know what documents you have. I found this one:


https://www.testworld.com/wp-content/uploads/service-guide-agilent-e4402b-e4404b-e4405b-e4407b-e4411b-e4403b-e4408b.pdf

In this document there is this block diagram for the tracking generator:




This is very useful. It shows that the course steps 8 dB steps are set by A2A3.
The LO oscillator is levelled by the A2A1 to +16dB
The fine control of the TG output level comes from modulating the level of the 600MHz A2A2J5 -11.5dBm to -4dBm

I don't know what instruments you have to confirm the signal levels.

But I suggest setting the SA to a low frequency say 10MHz and zero span. This will cause the TG output frequency to be fixed and easy to measure with a scope. Terminate in 50 .

You could measure the output at A2A2J2. As you change the TG output level the signal should change in amplitude by 8dB (2.5x) and then go back to times 1.

Once you have measured the output of A2A2J2 you can reconnect the A2A3 attenuator and measure the output again. This will confirm the attenuator is working. It seems be doing the 8dB steps, but it may have a high insertion loss.

After this it is measuring the A2A1J1 which is the LO drive at 3.9 to 6.9 GHz and 16dBm. This is going to be hard without another SA or a power meter.

You should also measure the signal level on the 600 MHz signal at A2A2J5.

If you don't have a suitable instrument, consider a power detector demo board.

Regards,
Jay_Diddy_B

[smgvbest]

Probably I should say I have the following equipment (test probes are where I am short) and limitations

My lab is not accessible at the moment so I'm not near the equipment, I am pulling things out as I need them
*Oscilloscope  Rigol 100Mhz
SA   *Rigol 815TG ((1.5 and most accessible) and Agilent E7495A (2.7Ghz)
*Tekbox Probes (near/far field)
2.5 Ghz Active Probe (untested, got off ebay)
*U2001H (USB 6GHZ)
HP E4418B w/ E4412
*Nano VNA

Those I starred are easy to access and move to the E4407/TG for testing

Note:   I did Terminate A2A2J2 with 50ohm and checked attenuator settings and had same "Source Unlevel" message, if it helps but seeing as it's measured before the A2A3 that may be expected?

it sounds like plan is
1. I need to re-assemble the TG and verify if "Source Unlevel" is at all frequency's (initial test was at 3Mhz-3Ghz) or some as well as the attenuator values
2. check 600Mhz in to TG at A2A2J5 (note SMB not SMA)
3. Check RF out at A2A1J1 (zero span and set based on finding in #1)
4. Make @analogRF RF Probe.
5. figure out how to measure Yellow section,  will likely have to make or find some extender cables to be able to access the board without the shielding on.  shouldnt' be hard. just need to do it

there's more I'm sure but a start

@analogRF
A2A1P1 to A2A2J1 Cable sends back ALC_MON to the A2A1 Driver board

I have both the service guide and CLIP but CLIP only shows the A2A1 board, 

Attached is where I mention the ALC_MON from the A2A2 board is feed back to the A2A1 board and into OP AMP to create the Unlevel Signal