Wavetek 288 SineAmpl Error and Calibration

[TinkerGDS]

I recently acquired a Wavetek 288.  After waiting 20 minutes for warm up and pressing the Calibration button the display shows ErrSineAmpl.  I don't have any history of the 288 so I don't know when it was last used, when it last worked, or why it was sold.  I've confirmed the Amplitude is definitely off.  I have a copy of the manual and decided to start the Performance Check just to check the entire unit and make appropriate adjustments.  The manual refers to Appendix A which apparently is the Performance Check Table and Record information.  Appendix A is not included in the PDF copy of the manual.  Rather than guessing at the settings and adjustments, I thought I would ask here, if anyone has a copy of Appendix A for reference.

[wn1fju]

You might take a look at the Army version (SG-1288) of the maintenance manual:  http://radionerds.com/images/a/a4/TM_11-6625-3198-40.PDF

On my unit, I had similar errors.  I reseated all the boards and the errors changed.  I then pulled out the RAM backup battery for a few minutes, re-installed
it, and then ran through the calibration (since the cal constants were lost).  If I recall, it took a few tries of the calibration before the thing settled down.  It
has been working flawlessly since.

[TinkerGDS]

Thank you.  That helps tremendously.  The Military version of the manual is much more detailed.

My version is the one that seems to be available on just about every manual site.  Yesterday, I reorganized all of the schematic pages in my PDF version in the order they should have been.  That will help me if I need to get to the board level for a repair.  Is there a place I can publish my version so others can take advantage of the correct order of pages?

[tautech]

Thank you.  That helps tremendously.  The Military version of the manual is much more detailed.

My version is the one that seems to be available on just about every manual site.  Yesterday, I reorganized all of the schematic pages in my PDF version in the order they should have been.  That will help me if I need to get to the board level for a repair.  Is there a place I can publish my version so others can take advantage of the correct order of pages?

Here maybe, how big is the file ?
Otherwise there's plenty of file repository websites online.

[TinkerGDS]

The file is 8.8MB.

[TinkerGDS]

The military manual definitely helped understand some of the adjustments.  I've finally been able to get through a Power On Calibration successfully now.  However, when I test the amplitude it is off about 6dbm.  I've started at the beginning with the sine wave adjustment.  When I connect my test gear per the diagram on page 2-55 I do not see any wave form that I can adjust peaks on step 6.  In fact, it looks like a pretty smooth sine wave, so I'm wondering if there's an error in the manual or if I'm overlooking something.  Were you able to get through that adjustment?

[wn1fju]

My reading of page 2-55 is to minimize the distortion of the sine wave.  It requires a distortion analyzer (something like an HP 332A or HP 334A is shown in the figure) and I'm surprised you have one, since it is not a common piece for people to have.  Anyway, I do have one and I'm getting pretty much what the
procedure describes.  As I adjust A7R33, the residual distortion (coming off the analyzer)  moves off of its minimum.  And A7R97 makes the resulting
peaks slightly asymmetric as the manual implies.  The ideal waveform in this procedure is essentially very low amplitude noise.  That means the distortion
of the 288 is good (since you are looking at everything but the 288's sine wave).

Your 6 dB amplitude error is curious, however.  Are you sure you have the output impedance matched?  In other words, if you hook up to the 288 with the
50 ohms selected, the amplitude as visualized on a standard 1M impedance oscilloscope will be exactly twice what is desired - i.e., exactly 6 dB.  You need
to have a 50 ohm terminator at the scope in order for the scope to show the voltage that is dialed into the 288.

Similarly, as you run through any adjustment procedures, you have to be mindful of whatever impedance they want at the measuring equipment.  Typically
they assume very high (>10M) impedance for things like voltmeters; but for scopes, it could be 50 ohms or 1 Mohm.  All depends upon the context.

And be careful of the calibration procedure while fiddling with the pots.  While playing with mine just now, I inadvertently exited the cal procedure incorrectly,
so maybe I corrupted some values.  The output was suddenly off a bit (1.1V for a 1V setting).  I did wait the warm-up time and let the thing do a full
auto-cal, and it seems fine now.

[TinkerGDS]

I was using 50 ohm impedance on my HP8903B Audio Analyzer with a 50ohm Feed Through at the input.  The 8903B has a monitor output that I connected to my scope's channel 1.  I have the 288 sync out to the scope's trigger in.  The scope is set to trigger on that signal.  The 288 output is set to 50 ohm unbalanced.  Since everything was set to 50ohms I set the scope to 50 ohms too.

I have the 8903B set to measure distortion with the reading in db.  Following the manual I set the 288 knob to a distortion level less than -50db.  I see the ripple, finally.  I've adjusted R33 and R97 to get the smallest ripple effect.  I think I've adjusted the ripple to have the lowest level.  Right now the delta between the p-p of the ripple is about 135mv.  I don't know if that's ok or not.

Next I adjusted R64 per page 2-56 for a near zero read on the multimeter.

After that I adjusted the amplitude since I adjusted R64.  This is where things get tricky.  I have three multimeters and I used each one to confirm 7.071Vrms.  I'm a little unsure about the rms part.  I'm assuming my Fluke 289 (TRUE RMS multimeter) is really displaying RMS.

After all of those adjustments, I connect the output of the 288 (still at 50ohms unbalanced) to my AC millivolt meter.  That meter has a dbm scale.  So I assumed if I enter -5dbm in the 288, I would see -5dbm on the AC millivolt meter.  But it shows -10dbm.  When I change the 288 output to 600ohms unbalanced with -5dbm the AC millivolt meter shows +1dbm.

Since this is the beginning of the calibration process I have not proceeded to the other steps as I'm not really sure if there is something I've overlooked.

[TinkerGDS]

Here are two shots of the results of the ripple adjustment from Page 2-55.

[wn1fju]

OK, it sounds like you know what you are doing.  I think you've exhausted my expertise on the subject, however.

All I can suggest is to run through the procedure.  But like I said before, when I was trying to answer some of your earlier questions, somehow my unit
got uncalibrated, possibly due to me incorrectly exiting the cal procedure.  Basically I cycled the power, let it warm up, and then ran the auto-cal procedure. 
That seemed to fix my (temporary) amplitude problem (although I was only off about 2 dB).

And if I recall correctly, when I first got my unit, I had to repeat the auto-cal procedure a few times before the thing stopped throwing errors or telling me
that a cal was required.  Didn't try to track that down because it eventually all worked fine.

[TinkerGDS]

I have confirmed the Amplitude Accuracy Test.  The amplitude is perfect, but it is entered in Vp-p not dbm.  Is it possible that the translation between Vp-p and dbm is off?

[TinkerGDS]

I appreciate your help.  I also had many calibration errors that ultimately went away after a few auto-cals.  The pots had to be changed quite a bit to get to a calibrated state.  I just don't know if the dbm display is somehow off.

[wn1fju]

At least on my unit, the amplitude display values seem correct.  For instance 0 dBm reads 0.223 rms and 0.632 peak-to-peak.  Similarly +5 dBm reads
0.397 and 1.12, and -20 dBm reads 0.022 and 0.063 respectively.  With a 50 ohm termination on the scope, I'm pretty much getting those readings
modulo small scope amplitude inaccuracies (using 50 ohm unbalanced on the 288).

It's almost like there is some confusion in your unit as to which source impedance is selected.  Is it referencing dB to something other than 50 ohms and/or
1 mW when selected?  Was this some sort of hidden menu selection or unknown option?  Don't know.

[TinkerGDS]

I decided to analyze the attenuator circuit yesterday.  I have an HP8662A Signal Generator that had a similar problem where it skipped the 5db attenuation.  It would go 1, 2, 3, 4, 9, etc all the way up.  In that case it turned out to be a bad pin on the 5db attenuator cable.  I repaired the pin and the attenuator started working.

I thought maybe a relay was stuck in the on position of the 288.  When looking at the manual and schematic for the 288, it explains that the Ladder attenuator was binary.  The schematic shows a 10 bit binary architecture.  A quad CMOS switch DG211CJ provides switching for the first four bits 1, 2, 4, 8.  Another IC provides a signal for six relays for the other six bits 16, 32, 64, 128, 256, and 512.  My scope has a 16 channel digital port, so I connected each channel to the 10 pins providing the attenuator switch signals.  I really expected to see a true binary TTL display - at 0 db all signals off, at one db 000000001, at 2 db 0000000010 and on.  It didn't work that way at all.  The signals I saw were pretty random.  I realized the attenuator pattern may not have been in dbm values, but probably in voltages including some decimal places.  I switched over to Vpp and increased the signal .001 Vpp at a time.  It made no difference.  The TTL signal still does not look binary to me.

I've read both manuals cover to cover trying to find anything that would result in a 5db increase in amplitude in all settings - like a calibration offset option or something like that.  There is nothing.  My unit is clean, spotless, and no parts have been replaced. There is no evidence of anyone ever tampering with any components.

[wn1fju]

This is all very confusing to me.  Set the 288 to 50 ohms unbalanced output and hook up a scope with a 50 ohm termination.  Toggle the 288 amplitude
button until it reads rms and punch in 0.223.  Verify you have that on the scope (0.632 Vpp).  Are you saying that

1) when you toggle the 288's amplitude button to dBm, the amplitude on the scope stays the same, but it does not say 0 dBm on the 288?,

or are you saying that

2) the 288 says 0 dBm but the amplitude on the scope changes (by 5 dBm)?

You're absolutely sure whatever meter you are using to read dBm is correct (verify with another signal generator)?  A lot of AC voltmeters use different
references than 1 mW at 50 ohms.  Just grasping at straws at this point!

[TinkerGDS]

I made the connections and performed the tests as you suggested so we're comparing apples to apples.  I had an eye-opening experience.  The scope showed 640mVpp.  When that appeared to work, I connected the 288 to my ACMillivoltmeter, and my multimeter.  Again, they showed incorrect values.  I then added the 50 ohm feedthrough, and the results were CORRECT!  My ACMillivoltmeter has a 600 ohm dbm scale, so I added the 600 ohm feedthrough to the cable and again the results were correct.  I double-checked the output when using the 288 dbm setting and it's off only .08 dbm.  At 0 dbm I see 0.08.  Accetable, right?  I had no idea I needed to use the feedthrough when performing the measurements.  Is the 288 designed to require a feedthrough on every connection?  I planned to use the 288 when making electrical adjustments to the audio equipment I am restoring.  Especially cassette and reel to reel tape decks.  They require precise calibrations.  Does this mean I need to connect a 600ohm feedthrough inline whenever injecting a signal into the tape decks?

[wn1fju]

Yes, I suspected from the beginning (my earlier post) that it might have been an issue with mismatched impedances.  Although I am still a bit confused by
your experiments.  So here's what I think you should be seeing.....

Most signal and function generators have a source impedance of 50 ohms.  The voltage that the user sets on them (and thus what is displayed on the unit)
usually assumes a 50 ohm load.  If the load is anything but 50 ohms, a voltmeter will show a value that does not match what is shown on the unit.

For instance, if you dialed 0.223Vrms on your signal generator, you would see on an AC voltmeter (with megohms of input impedance) the following:  0.223V
if you use a 50 ohm termination, 0.446V if the line is left unterminated, and 0.446*(600/650)=0.412V with a 600 ohm termination.  There is nothing wrong
with either the generator or the voltmeter.

The confusion becomes apparent when you have a piece of test equipment that was specifically designed for audio applications.  For historical reasons that
I can't remember, 600 ohms is the usual standard, not 50 ohms as it is in the RF world.  This confusion manifests itself when one tries to use dBm.  Whereas
a 50 ohm system equates 0.223Vrms with 0 dBm, a 600 ohm system equates sqrt(600*0.001)=0.775Vrms with 0 dBm.  A lot of my meters actually
have "0 dBm = 1 mW at 600 ohms" printed right on the meter face.

So, returning now to the 288 set to 50 ohms unbalanced.  If you were to select 0 dBm and feed an AC voltmeter that uses the 1 mW at 600 ohms dBm
definition, you would get the following:  1) If you terminated the line to the meter in 50 ohms, 20log(0.223/0.775) = -10.8 dBm;  2) if you terminated
the line in 600 ohms, 20log(0.412/0.775) = -5.5 dBm; and 3) if you left the line unterminated, 20log(0.446/0.775) = -4.8 dBm.   Note: all I did was
use the voltages I computed in the previous paragraph referenced to the 0.775V.

It can, of course, now get even more confusing if you are hooking the generator up to something other than a scope or meter with high input impedance.
For instance, if you were hooking up to a standard spectrum analyzer with 50 ohm input impedance, you would not use a 50 ohm termination.  All that
would do is parallel in another 50 ohm resistor, making the effective input impedance 25 ohms, and the spectrum analyzer would read incorrectly.

The moral to the story:  Volts are volts.  As long as the AC voltmeter has very high impedance, the voltage shown on the meter will always be correct.  It
will only match what is shown on the 288 if you have the line terminated in 50 ohms.  When you decide you want a result in dBm, you always have to ask
yourself what the meter's reference value is.  Otherwise you are guaranteed to find yourself confused.

As far as hooking up to audio equipment, if you don't know the actual input impedance of the equipment, I would tee-in a scope or voltmeter (with no
termination) to monitor the desired voltage you want.  Ideally you would want to match the source equipment to the load equipment (as it reduces things
like ringing or other distortions), but this may not be possible in every instance.  I would specifically stay away from dBm (or dBu, or dBV, etc.) unless you
compute the actual voltage it is supposed to be and then dial it in while watching the scope/meter.  Hopefully your tape decks tell you what voltage they
want, not dB values.

But I'm glad the 288 seems to be working!

[TinkerGDS]

Thank you for the detailed explanation.  I really appreciate that, and obviously needed it.  I am placing a copy of your response on my workbench wall for reference.  Unfortunately, the tape decks I work on all have different specifications for adjusting voltages.  Some are dbm, dbu, and dbv.  And most reference the settings in db, and some show the db value and millivolts in parenthesis.  I either need to keep a calculator near by, or make an Excel chart with the typical db ranges and millivolts as multiple columns for quick reference.

I'm still unclear about your comment:
"It can, of course, now get even more confusing if you are hooking the generator up to something other than a scope or meter with high input impedance.
For instance, if you were hooking up to a standard spectrum analyzer with 50 ohm input impedance, you would not use a 50 ohm termination.  All that
would do is parallel in another 50 ohm resistor, making the effective input impedance 25 ohms, and the spectrum analyzer would read incorrectly."

Does this mean I need to check each device (scope, millivoltmeter, handheld dmm, or bench dmm) before connecting the 288 to make sure I know which has high impedance and the feed through is not used?  Is this just something the 288 needs, or do I need to consider that with all of my signal and function generators?

[wn1fju]

It's a good idea to know what the input impedances are for any equipment you are hooking up to the 288.  As I said before, if the input impedance doesn't
match the 50 ohms output impedance of the 288, then not only could the equipment give you an incorrect reading, but also there may be waveform
distortion.  This rule of thumb applies to all function and signal generators, not just the 288.  It's all pretty much a simple case of voltage division between
the source impedance and the load impedance.

As an example, take any 50 ohm function generator and set it to output square waves.  Hook it up to a scope with a nominal 1 Mohm input impedance.
The amplitude on the scope will be double the setting on the function generator and, if your scope has high enough bandwidth (and the square wave
rise time is fast enough), you will likely exhibit some ringing on the pulse edges.  Terminate the cable to the scope in 50 ohms and the amplitude not only will
be correct, but the square wave will look a whole lot better.

In general, scopes nominally have 1 Mohm inputs (or 10 Mohm with a decent probe) and DMMs have at least 10 Mohm input impedances (the better ones
have several Gohms).  RF spectrum analyzers usually are 50 ohms, although some of the ones for cable TV are 75 ohms.  Usually the impedance is labeled
on the front panel.  A lot of auxiliary inputs, e.g., modulation inputs on a signal generator, are usually in the 1K to 10K range.  When in doubt, consult the
manual.  But knowing the impedances will keep you from making errors.  Believe me, I've been there myself many, many times!

[TinkerGDS]

When I turn on the 288 I get the "Cal. Required" message.  After I calibrated the unit, I realized the battery might be low or dead.  I replaced the battery with a new one and made sure it had more than 3v.  I spent a couple of hours doing the calibration again, since it was "lost" when I powered it off.  When I cycled the power I still got the "Cal. Required" message.  Now I have a feeling there's something wrong with the battery sense circuit since it is a brand new battery and I had calibrated it after installing the new battery.  I'm wondering if the Keep Alive LED should be ON when the 288 is powered off.  Or, is the LED only an indicator that the processor is working.

I'm wondering if the N-channel MOS (Q11) is bad as that is about the only component in the sense part of the circuit.  I will do more power off testing later today to see if the battery is powering the RAM or if the there's something else going on with the other part of the circuit.

[wn1fju]

I don't see how the light (CR6) can turn on unless Q1 is also turned on.  That can only happen if the thing is powered up.  So when off, the light is off.  When on, my CR6 actually blinks (as directed by the uP).

If I recall, I had to repeatedly run the front panel autocal procedure (after the initial 20 minute time-out).  Just keep running it until it finally passes.  Don't touch any of the adjustment pots in between.  If it keeps failing after a half-dozen or so tries, I don't know what to tell you....  Perhaps you have the adjustment pots
so out of kilter that it can't autocal.  Then you have to start from the beginning.

[TinkerGDS]

Thanks.  It certainly helps to have another unit to check your sanity.  I've gone through the manual calibration three times.  It looks pretty good except I can't get less than about -48.6 db THD on the Sine Wave adjustment (R33 and R67) Pages 2-54 - 2-56.  All of the other adjustments are good or great.  Mine will autocal each time without any error messages.  It just won't keep the settings after power off.

According to the manual the message "Cal. Required" would only be displayed if the battery were dead.  If low, "Low Batt" is displayed.  I never got a "Low Batt" message.  In fact, the old and new batteries are both good - more than 3Vdc.

Thanks again.  It helps to know what "Normal" behavior is.

[wn1fju]

With the unit OFF, I would trace the battery voltage from the battery, through the battery holder, through CR4 to U11 (the RAM that is powered by the battery).  I guess it is conceivable that CR4 is open, preventing the RAM from seeing the battery voltage, hence it loses its contents when the unit is powered down.  Meanwhile, the microprocessor looks like it monitors the battery through Q11.  So the microprocessor could be perfectly happy and not issue a low battery warning without even realizing that the RAM is going to lose its contents.   CR4 is a Schottky diode and should drop no more than 0.3V from the nominal battery voltage.  Most RAMs like about 2.5V to keep powered in standby, so you should be OK if everything is working.  If you are seeing proper voltage at the RAM,
perhaps the RAM is starting to go internally.  If you aren't seeing proper voltage, my guess is CR4, although I suppose a misbehaving CR5 might drag things
down also (less likely).

[TinkerGDS]

I tried the power off tests.  There IS 3.2Vdc at the RAM U11 Pin 28.  Since the processor checks Q11, I checked the signal from the source of Q11 to U3 Pin 5 (S5) while turning the power on.  Pin 12 gets the 3.2Vdc.  I don't really know how U3 works even after studying the datasheet.  I gather the processor U7 sends signals through A0-A2 to determine which switch (S1-S8) are being tested for a state?  I connected my scopes digital probes to A0, A1, and A2.  I saw A2 go high.  According to the truth table of U3, to get a switch signal on S5 DVM_BAT, A2 needs to be high, and A1 and A0 need to be low.  The enable voltage U3 Pin 2 is 4.9Vdc so that looks good.  I also have 11.9Vdc at U3 Pin 13 so that is good too.  I suspect U3 S5 is not closing.  Have I diagnosed this correctly?

[wn1fju]

Your explanation seems correct.  Although you shouldn't need to suspect U3 S5 is not closing.  Put a probe on the chip's output pin 8 and see if the battery voltage makes it through.  I don't know how easy it will be to catch the condition where A2, A1, A0 are 1, 0, 0, as I suspect that U3 might be toggling through many other selections - maybe you can set up some sort of triggering on that condition.   

What seems strange is that if U3 S5 is toast, why didn't the thing issue a low battery voltage message?