E4407B Tracking Generator Repair

[smgvbest]

I used a near Field probe,  calibrated the output thru a div/4 into my 1.5Ghz SA.  the reading was accurate to about +-1dB.

[zmetzing]

Could I ask what you searched for on eBay to locate one of these TG assemblies? I've got a E4402B arriving soon, and I'd like to add this option. I bought it due to the DSP/Comms hardware boards, which are tough to find. I'm still looking for a broken E4407B at a good price.

[analogRF]

Could I ask what you searched for on eBay to locate one of these TG assemblies? I've got a E4402B arriving soon, and I'd like to add this option. I bought it due to the DSP/Comms hardware boards, which are tough to find. I'm still looking for a broken E4407B at a good price.

just look for the TG part number as per service manual (look for replacement parts tables)
but it is very very rare and hard to find.

In my case for my E4407 I found a defective E4402 with TG on it and I convinced the seller to sell it for $400
then I kept the TG which was working and repaired the SA which had unlock LO and also bad attenuator (both simple repairs)
and then sold the SA  . Interestingly the TG was retrofitted on that SA so the sticker in the back did not have it listed 

 

[dj9ss]

Alas, some weeks ago my E4407 TG has developed the same symptoms: "Output Unlevel" at >2GHz at levels above -3dBm.

I already did extensive investigations and replaced all parts which were easy to source, even playing with the various attenuators to see whether there is any effect, but to no avail.
The only parts left are the mixer and the mystery 1BR MMIC which I couldn't identify.

However, I doubt the issue is in the 768MHz path since neither replacing MGA-6 with MGA-6A (more gain) nor changing the 5dB attenuator did change the symptoms.
I'd also rule out the LO amplifier with the two HMC311 - reducing the 10dB attenuation didn't show any effect.

Accordingly, my current conclusion is that it's probably the HMC129G8 mixer.
Unfortunately this part is obsolete and not easily available anywhere, so I'm planning to replace it with HMC129ALC4 and an adapter PCB.

During my work, I compiled a set of schematics in Kicad. There are surely still a few errors, especially in the control board, but I'll attach the current version anyway.
Additionally, I accidentially popped the cover off the mysterious black cylinder labelled 0410-2110 and take a picture. I bet that's a crystal oven.

Cheers, Jürgen

[dj9ss]

I finally succeeded replacing the HMC129G8 mixer with HMC129ALC4 using a self-designed adapter PCB - see attached picture.
Alas, it works exactly alike the original HMC129G8, i.e. the "Output unlevel" message at >2.5GHz and more than roughly -3dBm hasn't changed.
I could improve situation by replacing LEE-29 in the output amploifier by LEE-39 which has a few dB more gain.
Now I'm only getting the "output unlevel" only with >0dBm.

Otherwise, I'm frankly out of ideas what could possibly be wrong.
There might be an issue with the unknown 1BR amplifiers or even with the 1GG5-4202 output limiter.
The limiter is available at some places for horrendous amounts of money (several hundred $$), so I'm very reluctant to consider replacement.

Some update about the 0410-2110 crystal oven:
the oven part can be simply pulled off, it's just plugged onto the board. Inside there's a 92,140MHz crystal, see attached pictures.