E4407B Tracking Generator Repair

[smgvbest]

Tracing it so far. It does not go directly to the RF side of the A2A2 board
Goes thru a analog switch which switches between a fet op amp and J3 which is not connected,  then to Q1 which I haven’t been able to find yet.  It’s a 6 pin can.   See Attached

Oye,   I just did a SUT,   I was not tracing TUNE on Pin5 I was tracing ALC_EXT on Pin10.   I don't' know where I got Pin10 from but pretty much ignore my previous post.  I'll go retrace Pin5 for Tune and see where it goes.

[analogRF]

Tracing it so far. It does not go directly to the RF side of the A2A2 board
Goes thru a analog switch which switches between a fet op amp and J3 which is not connected,  then to Q1 which I haven’t been able to find yet.  It’s a 6 pin can.   See Attached

Oye,   I just did a SUT,   I was not tracing TUNE on Pin5 I was tracing ALC_EXT on Pin10.   I don't' know where I got Pin10 from but pretty much ignore my previous post.  I'll go retrace Pin5 for Tune and see where it goes.

I kinda knew because from the pictures I had already seen pin 5 (I had mentioned it is pin 5 it in my post) goes directly under that connector...

[smgvbest]

So when you trace the correct PIN its pretty easy to find.
See attached

[seronday]

Sandra,
I have just started reading this topic.
Correct me if I have misunderstood one of your earlier posts, but I understand that you have the schematic drawings for the 3Ghz A2A1 Driver board but not for the 3Ghz A2A2 RF board.

I have a printed copy of the drawings for the 1.5Ghz version of the tracking generator A2A2 RF board which I may be able to try to have scanned if this will be of any use to you.   (There are 13,  A3 size sheets).

Regards.

[smgvbest]

@seronday

Thanks for the offer. 
It definitely could help if your able to get those scanned

Sandra

[analogRF]

So when you trace the correct PIN its pretty easy to find.
See attached

(Attachment Link)

Still I cannot find what that sot-23 package MC-0 is but it appears to be (probably) a two diode package and since pins 1&2 are connected
it is a back-to-back connected diode which connects the TUNE voltage (0 to 12V) to that VCO
it all makes sense I believe...

the schematic of this thing is pretty much all clear now. you need a probing strategy to check various points
during fail conditions. But it is pretty clear now what you should expect at different points

I see most people who have had TG issues in these ESA series it usually has been the IGG5 hybrid ....

[smgvbest]

So not a exact replacement if this is what wrong.  I found this the attached.

The pinout is very close, mainly NC pins on the side that do not appear on the device in the unit.

I will pull out the Rigol tomorrow and probe before and after that component with the TG set to zero span and 50Mhz.  at -9dB i do not get the Source Unlevel,  at -10dB I do. so the output will be well withing the range of the Rigol SA,   I can also probe the 600Mhz in and the VCO.   I will get that done soon with a poor mans rf probe.


Also found this on Keysights site
https://www.keysight.com/en/pc-1748995/millimeter-wave-and-microwave-gaas-limiters?nid=-33897.0&cc=US&lc=eng

@analogRF
you mentioned you where going to post a pic of your RF probe,  where you able to find the pic?

[analogRF]

So not a exact replacement if this is what wrong.  I found this the attached.

The pinout is very close, mainly NC pins on the side that do not appear on the device in the unit.

I will pull out the Rigol tomorrow and probe before and after that component with the TG set to zero span and 50Mhz.  at -9dB i do not get the Source Unlevel,  at -10dB I do. so the output will be well withing the range of the Rigol SA,   I can also probe the 600Mhz in and the VCO.   I will get that done soon with a poor mans rf probe.

this chip does not have any peak detector output for level control loop. the one in your TG does RPP and has level detector.
it's good strategy to probe before and after that IGG5 when the error appears and when it doesn't
also at that point I suggest checking the control voltage that goes to the HMC346 variable attenuator and see if it is changing or not
and also check the the PWR_LVL signal that comes from A2A1 to the connector pin 8 and also the ALC_MON at pin 6.
these are all DC so easy to probe and document what you see

[smgvbest]

this chip does not have any peak detector output for level control loop. the one in your TG does RPP and has level detector.
it's good strategy to probe before and after that IGG5 when the error appears and when it doesn't
also at that point I suggest checking the control voltage that goes to the HMC346 variable attenuator and see if it is changing or not
and also check the the PWR_LVL signal that comes from A2A1 to the connector pin 8 and also the ALC_MON at pin 6.
these are all DC so easy to probe and document what you see

Set for 50Mhz Zero Span,  -10dB was Level, 3dB was unlevel.
reading at Top for Level, bottom was unlevel

see photo

Also

VCO?  782.5Mhz
see photo

[analogRF]

1-what are these numbers? what does -57 mean?
2-what did you measure them with? I mean was it scope or a spectrum analyzer?
3-the most important thing is what probe did you use?

4-for VCO this is very strange! but we dont know what that pin is so please measure the VCO
at the RF input to the mixer (pin 3) and also more importantly measure the IF output
you can measure at pin 2 of mixer but better is to measure at the output of the big filter which is the input pin of the blue "LF3" filter

[smgvbest]

Everything is in dB on first photo
Measured with Rigol 1.5 ghz SA
scope probe. Still waiting on sma to make that diy probe you mentioned

Interested in relative values in this case.  Should show if amps working
And IGG5 has same in/out


For VCO I measured with just a 50 ohm lead first. Then probe.  Got same value except probe was lower dB

I’ll check areas you mention.

[smgvbest]

1-what are these numbers? what does -57 mean?
2-what did you measure them with? I mean was it scope or a spectrum analyzer?
3-the most important thing is what probe did you use?

4-for VCO this is very strange! but we dont know what that pin is so please measure the VCO
at the RF input to the mixer (pin 3) and also more importantly measure the IF output
you can measure at pin 2 of mixer but better is to measure at the output of the big filter which is the input pin of the blue "LF3" filter

#4,   Pin3 is 182.5Mhz, 600Mhz Pin4  782.5Mhz at input/output of LF3

[smgvbest]

HMC346

3dB unlevel, then -10dB reading level
all vDC

Code: [Select]

V1 = 0.181/-1.7
v2 =  -4.191/-2.48
I  =  -0.585/-5.85

[analogRF]

Everything is in dB on first photo
Measured with Rigol 1.5 ghz SA
scope probe. Still waiting on sma to make that diy probe you mentioned

Interested in relative values in this case.  Should show if amps working
And IGG5 has same in/out


For VCO I measured with just a 50 ohm lead first. Then probe.  Got same value except probe was lower dB

I’ll check areas you mention.

you mean regular scope probe with SA?? that will kill all the signal and significantly disturb the signal path whenever you touch it
no need for sma connector, just solder a 450 or 470 ohm resistor to the center of a piece of coax like rg316 (or a rigid coax and then solder the rigid coax to a longer regular coax like rg316 in order to have a solid probe head) and then solder a piece of short wire to the shield of the coax (at the tip where you had the 450 ohm) to use it as your ground lead and your are good to go with 500ohm impedance you are good
use 20dB level offset on your SA in order to read the amplitudes correctly

something that is a bit strange is that between -10dB and +3dB setting the values dont change much. This could be due to attenuator section being
switched in/out so the levels at the TG out change little or it could be symptom of your underlying problem. based on the attenuator that you have at the output and also the range of TG level settings, you should be able to figure out what level at the output of TG must exist to have XdBm
at the front panel connector. Normally (not necessarily) for anything above 0dBm there should be no attenuator switched in but for -10dBm there is perhaps a section switched in like 8dB for example then your BITG must produce -2dBm at output.

anyways, if the closeness of the two measrements are symptom of underlying problem, then since the ALC loop is closed on the A2A1 board (the PWR LVL DAC) then i suggest check the control voltage the comes to the HMC346C8 chip. see the datasheet, there is an opamp on the other BITG board related to this chip. and also the voltage coming from PWR_LVL is 0 to -11V according to A2A1 schematic so it must be transformed into 0 to -2.5V for HMC346 and that is done in the other BITG board.
so you need to do some DC measurements. maybe there is bad opamp on that board!
also see if the PWR_LVL oltage coming from A2A1 is ok  (for example it is not stuck at a value)

[smgvbest]

I did check the Bias on the MCL29 which read 3.57vDC and the MCL59 which was 4.7vDC

at the front panel connector. Normally (not necessarily) for anything above 0dBm there should be no attenuator switched in but for -10dBm there is perhaps a section switched in like 8dB for example then your BITG must produce -2dBm at output.

at the -10db it says the 8db attenuator is active. 

I will have to trace the board to find the op amp the pwr_lvl goes to.

[analogRF]

I did check the Bias on the MCL29 which read 3.57vDC and the MCL59 which was 4.7vDC

at the front panel connector. Normally (not necessarily) for anything above 0dBm there should be no attenuator switched in but for -10dBm there is perhaps a section switched in like 8dB for example then your BITG must produce -2dBm at output.

at the -10db it says the 8db attenuator is active.

I will have to trace the board to find the op amp the pwr_lvl goes to.

good, so we know that at +3dBm the TG must output 3dBm (no attenuation after that) and at -10dBm the TG should output -2dBm (only 5dB jump)
your numbers did jump but not by 5dB between the two cases.

The control voltage for HMC346 clearly show the controller has gone into saturation in order to minimize the attenuation to retain the 3dBm output but has failed. HMC346 control voltages must always remain between 0 and -2.5V and the maximum tolerable range (before killing the chip) is +1V to -5V. So for +3dB
the controller is driving the HMC346 very hard to lower its attenuation.
In my opinion this is probably just a symptom and may not be the actual cause of the unlevel issue. For example, still IGG5 may wrongly report a low level and therefore the controller loop will drive the HMC346 into saturation because of that. But it also could be the source of problem, I mean maybe the DAC on A2A1 is giving out wrong voltage or the opamp circuits driving V1,V2 on HMC346 have problem.

it would be useful to check the ALC monitoring signal that goes out from TG to A2A1 board in the two cases and some levels in between to see how it changes

[seronday]

Sandra,
I have just started reading this topic.
Correct me if I have misunderstood one of your earlier posts, but I understand that you have the schematic drawings for the 3Ghz A2A1 Driver board but not for the 3Ghz A2A2 RF board.

I have a printed copy of the drawings for the 1.5Ghz version of the tracking generator A2A2 RF board which I may be able to try to have scanned if this will be of any use to you.   (There are 13,  A3 size sheets).

Regards.

I have managed to scan these drawings.  Some of the scans may not be very good as the printed copies are very feint.
There is a total of around 8Mb so I will split the files across a couple of posts.
Also included is a scan of the 1.5Ghz A2A1 Interconnect drawing for reference.

Hopefully they will be of some use.

Regards.

First lot of files.

[seronday]

Second lot of files.

[analogRF]

Sandra,
I have just started reading this topic.
Correct me if I have misunderstood one of your earlier posts, but I understand that you have the schematic drawings for the 3Ghz A2A1 Driver board but not for the 3Ghz A2A2 RF board.

I have a printed copy of the drawings for the 1.5Ghz version of the tracking generator A2A2 RF board which I may be able to try to have scanned if this will be of any use to you.   (There are 13,  A3 size sheets).

Regards.

I have managed to scan these drawings.  Some of the scans may not be very good as the printed copies are very feint.
There is a total of around 8Mb so I will split the files across a couple of posts.
Also included is a scan of the 1.5Ghz A2A1 Interconnect drawing for reference.

Hopefully they will be of some use.

Regards.

First lot of files.

This one is a completely different beast. Has no resemblance to the 3GHz version.

[analogRF]

HMC346

3dB unlevel, then -10dB reading level
all vDC

Code: [Select]

V1 = 0.181/-1.7
v2 =  -4.191/-2.48
I  =  -0.585/-5.85

change the level setting from -10dBm to 3dBm, 1dBm at a time and at each step monitor these signals (all DC) : the ALC monitor that goes from TG to A2A1. The PWR LVL that comes from A2A1. and V1,V2 on HMC346

[smgvbest]

change the level setting from -10dBm to 3dBm, 1dBm at a time and at each step monitor these signals (all DC) : the ALC monitor that goes from TG to A2A1. The PWR LVL that comes from A2A1. and V1,V2 on HMC346

I'm at a conference so been a little slow getting things right now,   will get them soon as possible
 
Ok, got them (see attachment)

Notes:   all reading done at 50Mhz, Zero Span.   Did one column at a time thru -10 to 3dBm
Also read the P15BITG, N15BITG and P5BITG,  the supply rails

RED Cells is when it went UNLEVEL,   but by time i got to ALC_MON/PWR_LVL the UNLEVEL state moved down to -1dB
BLUE Cells have variability to them, not a fixed value

[analogRF]

these numbers look pretty consistent with each other (except one thing that I will mention below) and i cannot see which part might be the main culprit. which part is failing that causes the ALC loop go into saturation. maybe someone else can draw a better conclusion

however here is my take and my recommendation
1-the PW_LVL signal coming from a DAC on A2A1 was supposed to be between 0 and -11V (says on the schematic) here it does not go below -9.7
it does not necessarily mean anything and it could have been designed that way because for example V1 should not go too much positive but still
it's worth to check if that circuit (PWR_LVL) is ok both on A2A1 and also on the TG board because that signal eventually generates V1 (so what is between V1 and PWR_LVL)

2-The only inconsistency that I see in your measurements and that may be the path to the root cause is that all numbers are and must be similar (very close) for -2dBm and -10dBm settings. Because the TG has to generate -2dBm for both cases. However there is ONE voltage which is very different between the two cases and that is the ALC_MON. In fact between -3dBm and -2dBm it suddenly jumps -1V which never occurs anywhere else. At -10dBm where TG must be generating -2dBm the ALC_MON correctly stands at about -0.17.

What happens to ALC_MON at -2dBm and above looks like the TG OFF state. Since I believe the ALC_MON is coming directly from the IGG5 hybrid (please confirm whether this is right or not) the issue may be the IGG5. Or maybe a supply voltage to IGG5 if it has one?

3-Your 5V rails looks a bit low. Is it clean? did you check it when the failure occurred?

4-since the failure moves to lower amplitude as time passes, it is probably affected by heat. Why dont you spray some freeze spray (aim very locally) at different components and see if the error goes away. starting with that IGG5!

[smgvbest]

these numbers look pretty consistent with each other (except one thing that I will mention below) and i cannot see which part might be the main culprit. which part is failing that causes the ALC loop go into saturation. maybe someone else can draw a better conclusion

however here is my take and my recommendation
1-the PW_LVL signal coming from a DAC on A2A1 was supposed to be between 0 and -11V (says on the schematic) here it does not go below -9.7
it does not necessarily mean anything and it could have been designed that way because for example V1 should not go too much positive but still
it's worth to check if that circuit (PWR_LVL) is ok both on A2A1 and also on the TG board because that signal eventually generates V1 (so what is between V1 and PWR_LVL)

2-The only inconsistency that I see in your measurements and that may be the path to the root cause is that all numbers are and must be similar (very close) for -2dBm and -10dBm settings. Because the TG has to generate -2dBm for both cases. However there is ONE voltage which is very different between the two cases and that is the ALC_MON. In fact between -3dBm and -2dBm it suddenly jumps -1V which never occurs anywhere else. At -10dBm where TG must be generating -2dBm the ALC_MON correctly stands at about -0.17.

What happens to ALC_MON at -2dBm and above looks like the TG OFF state. Since I believe the ALC_MON is coming directly from the IGG5 hybrid (please confirm whether this is right or not) the issue may be the IGG5. Or maybe a supply voltage to IGG5 if it has one?

3-Your 5V rails looks a bit low. Is it clean? did you check it when the failure occurred?

4-since the failure moves to lower amplitude as time passes, it is probably affected by heat. Why dont you spray some freeze spray (aim very locally) at different components and see if the error goes away. starting with that IGG5!

I will check these item out
on the IGG5,  I did check and I see no bias/dc at it's input so if it's  providing the level information I'm not seeing it.
I do have freeze spray and will target different IC's starting with the IGG5.   
I'll also check the P15/N15/P5 without the TG attached so see if it's still being pulled down.

[analogRF]

basically IGG5 should not need any DC bias, it has a peak detector and should generate a DC proportional to the power level
and that voltage after some amplification and conditioning by opamps or transistors will become ALC_MON

w.r.t. supplies, I think only the 5V rail seems a little low. check for any ripple/noise specially when the error happens and also for current draw.

[smgvbest]

basically IGG5 should not need any DC bias, it has a peak detector and should generate a DC proportional to the power level
and that voltage after some amplification and conditioning by opamps or transistors will become ALC_MON

w.r.t. supplies, I think only the 5V rail seems a little low. check for any ripple/noise specially when the error happens and also for current draw.

when I say no DC/BIAS I mean there is no DC of any kind at the input to IGG5.