Topics

Transformation TR503 into TR502


Attilio
 

Hello everybody,
I am writing about the attempt to transform my TR503 into TR502 to use it with the SA 7L13. I started by replacing the 55 MHz crystal with a 52.5 MHz one (thanks James) and after some testing I added a 180 ohm resistor in series with R412 on the "Iosc" circuit to bring the oscillator to 2095 MHz (thanks to Miguel and Ed). Then I re-calibrated the four-cavity filter to 2200 MHz. Now it works quite well up to a BW Resolution of 300 Hz (on Max span the amplitude variation is in the 2 dB), but when switching the resolution at 30 Hz the output signal from the TR503 drops by - 30 dBm and it is not possible to bring it back to normal level by acting on the "Tracking Adjust" of the TR503.
At the moment I don't know what to do, could it be the four cavity BPF that is not calibrated precisely?

Thanks for your help.
--Cheers
Attilio


Ed Breya
 

I mentioned earlier, that you should look at adjusting the new XO center frequency to get the tracking range to line up better. Also check the 7L13 calibration, which may include adjustment of its XO. Between the two, you need to get them to line up close enough. It's unlikely that the actual TG output drops at narrower IF BW - it's just the SA's response to it dropping down the IF filter skirt, on one side or the other. The narrower the IF BW, the closer the XO's need to match. My original post about this should be somewhere in the original thread.

Ed


Attilio
 

Thanks Ed,
I opened this new post because the other one was a bit too long and dispersive. I re-calibrated the four cavity BPF filter using the nanoVNA V2_2 by connecting it to the 2nd LO input of the TR503 and the mixer output of the TR503 (disconnecting it from the directional coupler of the oscillator at 2095 MHz).
By reconnecting everything, initially I was able to get the resolution to 30 Hz to work, but after a while the signal decreased and it was no longer possible to bring the signal back to the correct level by using the "Tracking Adjust" command.
Now I should find out if the TR503 or the 7L13 derives.

--Cheers
Attilio


Ed Breya
 

As the XOs warm up, they drift a bit. Check any mods or tweaking of their frequencies after they're stabilized. You're in the home stretch now - almost to the finish line.

Ed


Miguel Work
 

Yes, I remember using the TR502 in lower resolution settings, the need to set fine frequency adjustment. Set to zero span first and move the fine adjust for max response.

-----Mensaje original-----
De: TekScopes@groups.io [mailto:TekScopes@groups.io] En nombre de Attilio
Enviado el: lunes, 22 de febrero de 2021 21:27
Para: TekScopes@groups.io
Asunto: Re: [TekScopes] Transformation TR503 into TR502

Thanks Ed,
I opened this new post because the other one was a bit too long and dispersive. I re-calibrated the four cavity BPF filter using the nanoVNA V2_2 by connecting it to the 2nd LO input of the TR503 and the mixer output of the TR503 (disconnecting it from the directional coupler of the oscillator at 2095 MHz).
By reconnecting everything, initially I was able to get the resolution to 30 Hz to work, but after a while the signal decreased and it was no longer possible to bring the signal back to the correct level by using the "Tracking Adjust" command.
Now I should find out if the TR503 or the 7L13 derives.

--Cheers
Attilio


Attilio
 

Thanks Ed,
in the next few days I do some tests after a warm-up period.
I checked the 2nd LO output from the 7L13 and it is stable at 2200.002 MHz.

--Cheers
Attilio


Attilio
 

Thanks Miguel,
I'll try with span zero first and move the fine adjust for max response.

--Cheers
Attilio


Attilio
 

Ed,
I did some tests and I think I understand that the four-cavity BPF filter and the 2095 MHz oscillator are pretty good. I found that by holding a finger on the junction between the 52.5 MHz crystal and the varicap diode the system also works with the 30 Hz resolution set on the SA 7L13 and seems stable. I tried to connect both a resistor (1 Mohm) and capacitors of various capacities to ground at that point, but to no success or causing the crystal to not oscillate.
Do you have any ideas for replacing the finger with a circuit solution that gives a similar result ?

Thanks for your help.
--Cheers
Attilio


Attilio
 

Hello everybody
Ed, I understood the matter of the finger on the oscillator at 52.500 MHz, putting the finger the oscillation frequency is lowered to 52.497 MHz and I can recover the track by acting on the "Tracking Adjust" command of the TR503.
I also found that the downward frequency drift is due to the SA 7L13, after about an hour from switching on the frequency no longer drift. The problem is that when the 7L13 has stabilized I can no longer position the track at the correct level because the "Tracking Adjust" control is already at the right end (52.4981 MHz), while to bring the track back to the correct position the frequency should drop at 52.4978 MHz.
I disconnected the 52.500 MHz oscillator from the PLL input and inserted a DDS generator into the PLL input, simulating the frequency of 52.49783 MHz (with the correct offset and amplitude) and so the trace is in the correct position and is stable with the resolution at 30 Hz on the 7L13.
Now I ask you should I lower the frequency of the crystal oscillator or the problem is on the oscillator at 2095 MHz?

--Cheers
Attilio


Ed Breya
 

Hi Attilio,

You're almost there. The 2095 MHz LO should be just fine - remember it's controlled by the crystal oscillator, as long as the PLL is working. So, it looks like you just need to lower the XO to that 52.497 MHz for the center of the tracking range. It's usually easier to lower a frequency than raise it, so that helps. Did you ever try adjusting that "mode adjust" C524 that I mentioned way back? You should check that, and the adjustment procedure - it seems to be needed to suppress spurious oscillations in the original, but since the crystal is entirely different now, it may act differently. It has not apparently caused any trouble though, right? No matter what, it should have some effect on the XO frequency, so give it a shot, and see what happens. Use your counter, so you can resolve it well, and see just what that C524 does. If it does indeed do some "mode adjusting," you may see instabilities or jumps in frequency at some conditions, which could be a problem. If it can't get the frequency down far enough, try adding a little extra fixed C in parallel, maybe 10 or 22 pF - use NPO or mica type for this. Just tack it in for now. Once you figure it all out, you can make it permanent, if there's room for it. If it's too tight, you can change C526 to a higher value instead. All you need to do now, is tweak that center frequency and tracking range to line up with the 7L13, hopefully just with C524, and make sure the oscillator starts up and runs properly in all conditions.

Ed


Attilio
 

Thanks Ed,
I tried to adjust C524 but it is already at the end of the stroke because the frequency increases, but does not decrease. I will try to add a capacitor in parallel to C526 around 10 pF NP0 and check how it behaves.
Thanks for the help.

--Cheers
Attilio


Attilio
 

Hi everyone,
Ed, I made countless attempts by adding NP0 capacitors in parallel to the C524 but without getting anything positive, then I added a 10 pF NP0 capacitor between the quartz junction with the varicap diode and the ground and now the tracking also works with 30 Hz resolution set to SA 7L13. The track is a bit unstable with Phase Lock in the ON position.
Now I have another problem, the track is not repositionable using the "Tracking Adjust" control from a resolution of 3 kHz to MAX and I think it is normal, but it is also positioned about 3 dB above the correct level, is it possible to adjust it?

Thanks for the attention.

--Cheers
Attilio


Ed Breya
 

Hi Attilio,

I'm surprised that adding some C across C524 didn't do it. First, let's make sure we're on the same page - I'm looking at the TR503 manual schematic of the XO section. C524 is the variable trim cap "MODE ADJUST," 9-45 pF, and C526 12 pF is in parallel with it. The node you mentioned where the crystal and varicap CR526 join, is bypassed by C522 0.47 uF (apparently) to the same cold ends of C524 and C526, which is the "ground" for the oscillator. The actual DC level is set by the Vbb reference from U520, so its ECL receivers U520A and U520B will be properly biased. U520 is set up for PECL (positive ECL) operation, so Vbb ends up at +3.6 V, as marked on the schematic.

The first question is what is C522 - 0.47 uF, or 0.47 pF? The label is kind of ambiguous. Other Cs in the area that look like bypass spots, have the same label, "0.47." It would make more sense if C522 is 0.47 pF (although that seems way low for something like this), because then it's clear how the feedack gets to the input. If C522 is indeed 0.47 uF, then it should effectively short out that end of the crystal to "ground" (Vbb), and no obvious feedback would be available at the front. Since ECL has very little gain and isolation from output to input, it could be some scheme where the feedback is internal, but I don't recall ever seeing this kind of topology. I suspect that there may be one or more errors in the schematic, or missing parts.

I just took a quick look at the TR502 manual, and the circuit looks the same, but shows a mod R522, unidentified value, between C522 and the crystal/varicap node. Now that would make more sense - C522 can be big, for bypassing the tune voltage, and R522 would isolate it so some feedback is there to sustain oscillation.

So, I think it would be good to check the possible different versions of manuals for both models, to get a better idea of what's actually in there. Then, double check that your added 10 pF experiments were truly at the right spots, across C524 or C526, and if there's any effect with 10 pF, but not enough, then try more, like 22 pF, and so on. For any shunt capacitors, the ground is the Vbb, not the system ground. You can try various experiments and see how it goes. If reasonable extra C across C524 really doesn't work, then I think the next best spot would be some C across the varicap diode, but it may need to be a lot less than 10 pF. If you can, find the specs for CR526, and see what its capacitance range should be - that will give some idea of how much extra tweaking is needed.

Ed


Attilio
 

Hi Ed,
so I confirm that the C522 capacitor is 0.47 uF (470000 pF).
In the diagrams there are several inaccuracies, but if you look at the drawings of the printed circuits they all show the resistor R522 (1000 ohm).
CR526 diode (MV1404) is called "Hyper-Abrupt Tuning Diode" the nominal capacitance is 120 pF (the range should be from 10 pF to 140 pF).
I put the 10 pF capacitor (NP0) between the crystal terminal (connected to CR526) and the system ground (basically the ground where the 55 MHz crystal container was connected).
I tried several capacitors in parallel to the 12 pF capacitor but without positive results, with 22 pF capacitance the oscillator by adjusting the "Tracking Adjust" stops oscillating, with 10 pF there is not enough adjustment, in any case the track is not stable.
What I try now?

Thanks for your help and patience.

--Cheers
Attilio


Ed Breya
 

The varicap diode is pretty high in capacitance, so 10 pF added is probably about right. Put the cap across the diode - not to ground. Ed


Attilio
 

Hi Ed,
I'll try tomorrow.
Thank you.

--Cheers
Attilio


Attilio
 

Hi Ed,
sorry if I wrote an inaccuracy, the 10 pF capacitor is not between the junction point of the crystal with the varicap diode and ground, but on the other terminal of the crystal towards ground.
However I moved the capacitor in parallel to the varicap diode and it is better. I am left with the problem that from 3 Khz to MAX resolution the track is about 4-5 dB higher on the SA display and I don't know how to bring it back to the correct position.

--Cheers
Attilio


Attilio
 

Hi Ed,
on the TR502 Tracking Generator manual I read this note "NOTE: If the TR502 is used with spectrum analyzer 7L13 the TRACKING ADJUST may run out of range at high resolution (3 kHz or less) setting because the IF in some early production 7L13 units may deviate a small amount from 105 MHz. If this occurs contact your Tektronix Field Office or representative to correct this situation. ".
Could it be related to my problem?
Even if it were I don't know if I could get an answer from Tektronix.

--Cheers
Attilio


Ed Breya
 

Atillio, the TG may be OK now, producing the right center frequency. You should be able to get the amplitude to peak at narrowest bandwidth, and drop off when the tracking pot goes either way from the peak. Once this is working right, nothing should change in the actual output signal level when you change the SA's IFBW. The amplitude response is then up to proper operation, and calibration of the SA's gain at each IFBW setting. I think each one is adjustable. First, be sure you're not the cause by running the sweep too fast for the BW, - the narrower it is, the slower you have to sweep. I think there's supposed to be a "UNCAL" warning if the SA setup is wrong, but not sure. Just be aware of the issue. If the amplitude response really is wrong or different between settings, then look at the 7L13's calibration.

Ed


Attilio
 

Thanks Ed,
yes, the behavior of the tracking generator is as you write, I have a maximum then the level goes down by turning the Tracking Adjust control left or right. I was careful to scan right so that the "UNCAL" light was turned off (for 300 Hz and 30 Hz resolutions I use the 7613 mainframe in STORE mode). At first I did not realize that the track must be adjusted with the Tracking Adjust control for maximum peak on the screen upwards.
If I understand correctly I should calibrate the 7L13 so that the track level is stable and in the correct position by varying Span / Div and Resolution from minimum to maximum.
Do you have any suggestions?

--Cheers
Attilio