开云体育

Hi all. I'm back with another amateur mistake in my audio lab. My 5L4N Spectrum Analyzer died from a short in the 5440 scope. I was going to start a calibration procedure on it using cable extenders. The extender must have slid down a little in the mainframe scope. When I turned it on it smoked, hissed and blew the mainframe fuse, and fried two tants next to the plugin sockets. Once i got the scope going again I realized the 5L4N was no longer displaying any specific frequencies. Even the comb output wouldn't display anymore. Nothing looks burnt inside which surprised me because I probably sent 200v into it! I do have a display but just the baseline noise like when there no input signal connected. The noise display responds to the attenuator and span adjustments.

It's taken a few weeks to get hold of the final version service manual and make new cable extenders with keys to keep it from sliding down in the slot. I verified the comb output was good on my HP spectrum analyzer. This is a really complicated unit. So not exactly sure where to go after verifying the internal power rails. +15, -15, and -5v are all good inside the unit and power is intact at least through the input stage.

I was thinking that the problem was at or after the mixer because the comb signal enters at the mixer. But as I'm writing this realize that can't be true because you have to jumper the comb output to the input bnc. So the problem could be as early as in the input stage.

Is there are method I could use to determine what is an appropriate signal between the main stages on the block diagram? How would you guys approach this.

Thanks, Rick

Rick,

I would troubleshoot it like any receiver. Put the 5L4N into manual mode to
stop the sweep, set the resolution bandwidth potentiometer to 3 kHz and then
inject a signal into the the input connector. Maybe something like 10 kHz
at -10 dBm. Then, follow the signal with an oscilloscope or a low frequency
spectrum analyzer down the chain until it disappears. I would pay
particular attention to the inputs and outputs of the first mixer. The
frequency tuning will need to be adjusted to about 10 kHz to set the local
oscillator frquency so that the it is offset by 250 kHz from the input
frequency to bring a signal into the 250 kHz IF. Also, there is a small
0.1A fuse on the input. Perhaps, that is blown?

Brian Henry

Brian, this sounds like a great way to start. Wouldn't that be great if it turns out to be the fuse. I'll know tonight.

Thanks

Brian, the fuse was okay. I chased a 10kHz signal to just before the L pad network controlled by relay K160. The signal is present on the negative side of the coupling cap C184 but does not appear at the center tap of T218. So checked the +15v and +30v at relay K160. They're good and the relay is receiving a switching signals from Vertical Gain Selector. I think either the relay doesn't make contact properly or some component in the L pad network is bad. I tried bupassing the K160 by jumping C184's negative side to R186 then from the other side of R186 to T218. That didn't bring it alive. I don't really understand what the resistors around the mixer balance adjustment are doing other than trying set the level up properly for the mixer.

The signal at C184 is severely clipped almost to a square wave. I noticed when I took the frequency of the test signal up to 100kHz it became sinusoidal again. But I think it is because my SG505 signal source is attenuated quite a bit at that frequency. In going back to your note to me I see you said test with about a -10db signal and oops, I was using a 0db signal. Maybe that's why it was clipping that late in the input amplifier. I had shut the bench down for the night or I'd go confirm that was the reason for the clipping. Although it seems like the unit would attenuate a 0db signal it without clipping.

I really appreciate your input. Thanks, Rick

Rick,

The 5L4N might be overloading with a 0 dBm input level. -10 dBm is the
highest value on the input attenuator.

The L Pad ahead of the first mixer inserts either 14 dB or 24 dB ahead of
the first mixer depending on the position of K160.

What does the LO injection look like and what does the output of the first
mixer look like.

The frequency response of SG505 should be flat to 100 kHz. You might want
to verify that it is.

Attached is the circuit description for the 5L4N. I hope that you find it
helpful.

Brian

I had a few minutes to check it with a lower input signal. I was overloading it. As long as I drop the input down to -10 to -12db the signal all the way down to the L-pad relay is undistorted. The signal only shows up at the entrance to T218 if the relay is in one position. It is very diminished in the other position. The other thing is I have lost the signal by the time I get to C210 at the end of the mixer in any position of the the relay. I was short on time and didn't have a chance to compare the dropped signal to the schematic. I'll do that tonight.

Rick

I've got a clean 260mV wave at the center tap of T212 at the same 10kHz as the injected signal. The input signal is -15db. As I mentioned earlier I only have this signal when the log/lin cam switch is in only one position. All other switch positions kill the wave to zero. I don't have a signal at the C210 and C213 junction. I'm not quite sure how to take other measurements from the mixer. Seems like the diodes are okay. T218 seems like it should have a signal on it on the way to C210.

Any more ides would be great.

Thanks

The LO input to the mixer is also dead. It comes into the mixer through a cable/plug and it's still dead while unplugged. There are several test points on the LO circuit . The manual has a few reference wave forms I can check. In summary, both the LO is dead and the mixer's output transformer isn't even passing the basic AF wave by itself. But the transformer isn't open. I'll dive into these over the weekend and see what I can find.

I decided to chase down the missing LO signal coming into the mixer. The oscillator signal is correct all the way through to U1040 output 9 at 10MHz. The output pin 12 of U1042A JK Flip Flop is divded by 2 with 5MHz on the output. However, output pin 8 of U1042B is still at 5MHz and at about have the voltage. It's suppose to dived by 2 agains down to 2.5MHz. Then it heads over to U1044. The output pin 12 of U1044 is dead. U1042 is very hot but the power rails are good. I say good. The 5v rail is about 4.85v through out the unit. I think U1042 maybe bad and even the decade counter U1044 could be bad. I guess I could pull these two chips and see if the 5v comes back up to a full 5v.

I found both chips at Very resonable.

I read that a double balanced mixer like this one suppresses the fundamental frequencies of the LO and AF frequencies and only leave the difference or additive frequencies to output. Now it makes sense that I had no main AF of 10KHz at the mixer's output. If the LO is dead there's no mixing and no other frequencies to output leaving it with nothing. So I feel good about switching gears to focus on the LO source problem.

I went ahead and pulled the decade counter U1044 out of the circuit and U1042B started properly dividing by 2 bringing the LO signal down to 2.5MHz. To be sure there wasn't anything else shorting out U1044 I lifted C1043 and unplugged P1040 (the final LO output signal) so the U1044 was some what isolated. U1044 still has no output. I suspect it has an internal short that is drawing too much current through U1042. The new U1044 should be here in a few days.

Turns out I had a decade counter on hand that met most of the specs. That did the trick and the unit is now working. The LO signal makes it to the mixer and the analyzer displays the spectrum. The remaining problem is that the -20dB gain setting is very low or not passing the signal at all. This is an original problem and not related to the short that blew out the decade counter. I'll chase this down in another thread. Thanks Brian for kicking me off in the right direction.

Rick