[SOLVED] Circuit for testing a x-tal

[hafrse]

Thanks again for your help, I am very greatfull for that and not usual to be lucky to get that help , it was little tricky to know that the cap was bad even though it gave correct value with my capacitance meter. It probably leaked as you proposed.

Best regards

 

[hafrse]

Hello,

Still one problem exsist, on instrument startup, the instrument executes test routines on the tune voltage of the 29 oscillator and counts to the processor to check if the oscillator is oscillating at its tune limits (-4V and +10v):
First it starts with +10v on the tune voltage, oscillator will not run, then it goes down to -4V , the oscillator oscillates, then it goes to its normal value 3.72V when the instrument is ready after start and oscillator still runs.
After the startup sequence, the instrument report the error message because it did not count when the tune voltage was at 10V.
My question here, why does the oscillator does not start when the tune voltage is at 10v ? does the unknown value of the rplaced bad c6 (100nf ?) cap hav any thing to do with that?
Thanks

 

[E-design]

The varicap capacitance value may be too low at +10V. Test the oscillator for operation at +10V tuning and increase the value of C4 over the diode until it starts to oscillate.

 

[hafrse]

Hello,

It is more complicated, it hasn't to do with the tune voltage, I took away the tune voltage and still the same problem.

The circuit attached is the same except the oscillator circuit in the red rectangle is from a newer version , the one I have is the hand drawn in my previous post.

Explenation:

There is a +15v swtiching signal for enabling and disabling the 29mhz to the rest of the system, this is because on higher span, the 29Mhz is not needed.
On startup, the switching signal is +15V to enable the 29 Mhz path to the system and the counter for 3 sec for diagnostic reason (and only here it fails!). where also the oscillator itself should start by power up independent of this switching signal, then it goes down to 0V, at this level, the oscillotor starts for some reason.
After the startup sequence I change the span in the instrument which means Switch signal +15v or 0V the oscillator works and does not stop as expected, did that many times.

There is something not stable here since switching from +15v to 0v it kicks the oscillation. Also. if I try to measure (by scope) on the cap between the oscillator circuit and the rest then the oscillator stops and need to restart the system.

Thanks

- - - Updated - - -

Update:

Took away the two 100 Ohm parrallel resistors to disable the +15V switch, the oscillator always starts.

 

[E-design]

Try and insert for testing purposes some series resistance (100Ω to 1kΩ) with the coupling output cap of the oscillator and see if it starts while the +15V is still present. It appears that something is loading down the oscillator too much for it to start up with the +15V applied to the switching circuit. The small resistance may be enough to reduce the loading to facilitate the initial start-up.

 

[hafrse]

Hello,

Did you meant a series resistor after the coupling cap or before the coupling cap ? Thanks

- - - Updated - - -

Update:

Inserted a resistor between the switching circuit and the coupling cap 820 Ohm, it made the oscillator never starts. It seems that the driver/switching circuit is matched to run with the oscillator.

 

[E-design]

Do you know if your switching circuit is the same as the later version shown in your attachment?

 

[hafrse]

It is the same as far as I know but will check that .

- - - Updated - - -

Checked it now, it is the same.

- - - Updated - - -

Also, When I took away the two 100 Ohm parrallel resistors to disable the +15V switch, the oscillator always starts/starts from the beginning as stated before and the good thing is that the diagnostic routine counter manage to count at startup even though the signal level is way under 0 dBm (measured at LO Out socket)! so may be to solve this problem is to add a delay timing circuit of about 4 sec. to cut the +15V while it is performing the startup diagnostics and then enable the swithching signal. Is it possible to do that with a simple transistor and RC circuit connected to a permenant 10V on the board itself ? or must I use a relay ?
Thanks

 

[E-design]

We can start modifying circuits to get it working but that only means we are ignoring the fact that there is still a problem somewhere as it worked correctly in the past.
If it was me, I will take a close look at the components in the switching circuit. You may have other leaky components that prevent this circuit from operating as it should and thereby affecting the oscillator. By looking at the schematic, the only connection to the oscillator is via the coupling cap. I don't see any other connection that can prevent the oscillator from starting up.
Maybe you can measure some DC conditions (using a storage scope if you can) at the point where the coupling cap connects to the switching circuit, so we can see what happens during the first few seconds?

 

[hafrse]

You are right, will give it a try, there are two coupling caps (c10,c11 in my drawing) in series where I can not measure there values since the value is very low, in the new design there is only one cap .01uF . I tried to replace them with 2 none smd 47pf caps, but the oscillator did not run at all.

I measured the DC conditions at startup where it connects to the switching circuit, on start when switch +15 is activated,dc condition is +3V then it goes down to 0V and the oscillator kicks on. After that I can alter the span so that the switch goes high again and the DC conditions is the same, +3V and back to 0V, oscillator will never stop.
As mesntioned before, if I try to probe before the two caps or betwen them the oscillator stops completely and I need to restart the instrument. It seems to be very sensitive there and If i take away one coupling cap then the oscillator will not run (or it stops directly when I probe). It seems that a part of the switching/driver circuit works as an output loadinbg for the oscillator (?). Thanks

- - - Updated - - -

update:

Replaced the two coupling caps with one 10pF, and it worked! I will do some more test, thanks again for your help!

 

[E-design]

In your design, they are very small because the tap of directly from the tank, and the loading impedance must be a high value. The new design has a buffer stage and maybe was redesigned due to sensitivity problems with the single stage. Good job!

I always used to tell people working for me: "I want you to fix it, not redesign it"

 

[hafrse]

Thanks, same to you! and very wise decision. It faild some of the self cal routines (29 LO out of range etc..) from the beginning since the value is unkonwn for these coupling caps, I tried diffrent vaules, 20pF made all the self cal routines pass! repeated them many times today, and does not fail a single step. Forgot to say, the instrument is a Tek. 2782