Electronic Test Equipment Qs

Knowing that electronics engineers and hobbyists constitute a greater than normal share of the Prius owner population, I am posting my questions here.

1. I'm interested in knowing what instrumentation you find especially valuable, particularly for analog electronics work. I'm amazed at the low cost on eBay for older, used test equipment manufactured by HP, Tektronix etc. As I am a hobbyist and not earning my daily bread with this equipment, I'm not interested in spending tremendous amounts of money on test gear.

2. I've owned an HP 275 MHz two-channel analog scope for > 15 years. What is the benefit of a digital scope?

I recently bought an 100 MHz HP universal counter (two channel) for $66 + $18 shipping and received it yesterday. It's amazing to think the successor Agilent product sells for $2,300, new.

I used my MFJ HF/VHF/UHF SWR analyzer to generate RF from 1.8 to 100 MHz and the counter measured all. It gave up at around 105 MHz. I also applied a signal to the second channel and the counter computed the ratio between the frequencies applied to the two channels.

3. I'm now trying to think of a way to calibrate the 10 MHz clock on the counter, without access to a lab frequency standard. My current thought is to pick up WWV in Fort Collins, CO on 10 MHz using a receiver, introduce the signal from the counter clock, and see if there is any noticeable beat between the two signals. Any comments on that?

I've also recently bought an old HP sine wave generator which operates from 10 Hz to 10 MHz. This uses a Wien bridge oscillator and should be arriving on Monday, we'll see if it works or not. Hopefully I can restore it to life, if not currently functioning.

Thanks for your comments!

 

Hi Jeff,

Yes, I've noticed those PC-based scopes, fascinating.

Unfortunately the WWV 10 MHz signal here is relatively weak compared to ambient daytime noise level so I couldn't use my first approach of comparing WWV vs. the counter clock signal directly.

My Kenwood TS-2000 transceiver reads frequency to 0.01 kHz and according to that the counter clock signal is at 10 MHz. So I'm pretty confident that the 10 MHz clock within the frequency counter is within 20-30 Hz and that will be good enough for my purposes (although the HP documentation allows you to set the counter within 1 Hz if you have a reliable reference to work with.)

Thanks for offering your suggestion regarding measuring local oscillator frequency. That suggestion assumes that the IF frequency is exactly per the spec. However, it seems quite possible that the IF could be off as much as 0.1 MHz given for example, a first IF of 55 MHz.

BTW, HP and Wikipedia say that the first HP oscillator design was based upon W. Hewlett's master's degree thesis. [ame]http://en.wikipedia.org/wiki/Wien_bridge_oscillator[/ame]
I am getting the HP 654A. Here's a used equipment vendor selling a tested and calibrated model for $300:
http://www.testelectronics.com/used/hp654a.htm
I paid $98 plus shipping for mine, we'll see if it works or not.

You'll have to decide what use you'll make of the scope. I don't know if you recall the 60's when Heathkit made 25" color TV kits for sale. One use of an oscilloscope back then would be to view the video signal, which required a vertical amplifier bandwidth of only 5 MHz.

Since my interest is primarily ham radio, I was interested in an oscilloscope that would have a relatively high bandwidth, in my case it is 275 MHz. I recently applied a 440 MHz signal to the channel 1 input to see what would happen. The scope showed a signal but the internal trigger would not lock on it and of course the displayed amplitude of the signal was considerably below what it should have been.

Yes, I've noticed that the digital scopes show frequency, amplitude, etc. displayed on the screen, which is quite cool.

What does "pre-trigger" mean?

 

Thanks, Richard. It would be great to have that, along with a two-tone audio generator available to evaluate the distortion in my SSB signal and see what happens when the speech compressor is used vs. no compression. At this point I have to rely upon the fact that no one has complained about my signal quality.

Another possible use for the analyzer would be to connect it to the output of the receiver's first IF so that you can see all signals allowed through the IF passband. That would enable you to quickly see where the strong signals are, for example, without having to tune to them.

Yes, I need to find a time when I can pick up WWV at 10 MHz with a reasonably steady carrier so that I can look at S-meter fluctuation to see how much beat exists between WWV and the counter frequency - since I can't hear an audio tone below 100 Hz or so.

Thanks, excellent suggestions.

Patrick Wong AK6C

 

OK, I will claim victory on this now. Here's the procedure I used:

1. I found a location where there's relatively little atmospheric noise and I can pick up WWV, 10 MHz, at around S8 on the Kenwood TS-2000 S-meter, using a 12 ft. length of wire. The HP counter produces a signal of S9 so the two signals are pretty comparable.

2. I chose the CW mode and set the filter narrow enough, at 200 Hz bandwidth, to focus on the carrier signal. I tuned to 10.00003 MHz to produce an audible tone on the speaker.

3. I hooked up my Fluke 87 to the audio output and set it to read Hz with the counter momentarily turned off. The audio tone ranges from 784.3 to 784.4 Hz.

4. I then turned the counter back on and after a little time to stabilize, the audio tone reads 782.2 to 782.8 Hz.

5. I turned the counter off to confirm that WWV continues to produce an audio tone frequency per #3 above.

I therefore conclude that the counter frequency is now within ~2 Hz of WWV, so I can claim the counter accuracy is around 2 parts in 10^7 (at least as of this writing. Who knows what will happen by tomorrow...)

It took considerable adjustment of the OSC ADJ slug on the counter to get that close. If the slug is tapped too hard, the frequency will jump around.

Thanks to all for your help and suggestions!

 

I'd like to ask the group for help here:

1. The HP 654A test oscillator produces a good signal across the entire range. However the output was slightly less than it should be. I removed the covers and found two electrolytic capacitors that were dead. Both were scorched, one had electrolyte leakage, and a wire actually had broken off the end of the other. They are related to the automatic gain control circuit.

2. After replacing the capacitors, I found that the generator now produces 30% more output than it should. This is partially because two transistors in the "amplitude control integrator" section had failed, probably related to the capacitor failures. I just placed an order with Mouser for replacements.

Further, there is a photocell module which provides isolation for the automatic gain control. The bulb within has failed.

The module contains a resistor which measures ~3k ohms. CLM5012 is the manufacturer's part number according to the HP service manual. The HP part number is 1990-0082. The manufacturer is not identified. I can't find this part on the web, so my idea is to cut the module open with a Dremel tool, replace the light bulb with an LED, and use electrician's tape to seal the opening.

Any alternative suggestions for the above? After I replace the two transistors mentioned above, I'll measure the voltage that would be applied to the bulb so that I can determine the value of the resistor to put in series with the LED. Right now, the voltage is ~20V, due to the failed transistors.

Those were great reviews, thanks for sharing.

I want to see whether I can fix this relatively simple HP test oscillator. If I can't then there's no hope that I would be able to fix a subtantially more complex device.

Yes. I have confidence that I can calibrate my frequency counter to a very high standard. Using that and my Fluke DMM, I have found that my oscilloscope is reasonably well calibrated with regards to the time base sweep frequency and the vertical amplifier volts/div.

Assuming that I can get the test oscillator fully operational, then I will go through the calibration procedure. As-is, it is within the spec for frequency.

 

Yes, I also decided that using an LED was a bad idea after playing around with one last night and thinking about the resistor required to control the current flow.

This morning I unsoldered the old photoresistor part. Unfortunately the leads that connect to the CdS sensor broke off, right at the epoxy used to seal the part.

I made a cut in the part around 1/8" in from the end containing the lamp and removed the old lamp.

I then went to RadioShack and bought a package of 5 CdS sensors of assorted sizes, as well as a package of three 6V light bulbs, in red, green and yellow.

I first tried using superglue to glue the yellow light bulb to one of the RadioShack sensors. Unfortunately the glue would not bind to the bulb glass or the sensor clear coating.

I then tried cutting off 1/8" off the end of the old part, on the CdS sensor side. To my amazement, I was able to remove just the epoxy and the thin aluminum case, exposing the CdS sensor leads sufficiently to solder extensions.

Then I stuck the RadioShack yellow light bulb into the other end of the old part, and used black electrician's tape to seal that end.

I soldered this repaired part back into the HP 654A test oscillator, and to my amazement it actually worked.

Hence what was needed to restore this oscillator to proper operation was to replace two obviously failed electrolytic capacitors and to repair that photoresistor assembly. (The neon light which shows power-on has also burnt out, and I ordered a replacement via Mouser.)

I also replaced one transistor yesterday because the voltages were way off, but later realized that was unnecessary and didn't help.

I set the power supply voltages and the oscillator feedback to the exact values as specified in the manual. I also set the output voltage to be exactly 2V peak-to-peak and 0.707V RMS, at 10 dBm (unbalanced 50 ohm load).

That output voltage is maintained across the entire oscillator range without adjustment being required, per my DMM (at lower frequencies) and my oscilloscope (for the remaining range up to the 10 MHz upper limit.)

The frequency calibration is much better than spec so I left it alone. The attenuator seems to work within calibration, given the limits of the oscilloscope and DMM. Its range is from +10 dBm to -89 dBm.

The only complaint that I have is that above 2 MHz, the oscillator frequency stability isn't perfect. There is a slight drift. I can see the counter showing the frequency moving one or two Hz every few seconds. Note that there is no spec for frequency drift.

Thinking about HF amateur radio equipment, the usual design spec is to have the VFO operate over 500 kHz or 1 MHz. Over that range the expectation is for the VFO frequency to be totally stable.

Since the top range of this test oscillator requires continuous tuning from 1 MHz to 10 MHz, I suppose it is asking too much for the oscillator to perform with high frequency stability over that entire range.

Thanks again to all for your support. It's nice to know that I can fix this level of equipment given the availability of the repair manual documentation and your help.

 

Yes, I would say that if the test oscillator is being used for frequencies of 1 MHz or less then a 30 min. warm up period is sufficient but if you want to use the top range of 1 - 10 MHz then 24 hours warm up would be helpful.

I've attached a few photos that show the two damaged electrolytic capacitors, the CdS photosensor assembly prior to my repair (which is marked VACTEC), as well as photos of the oscilloscope showing the test oscillator's 2.5 MHz sine wave output at 2V peak-to-peak into a 50 ohm load, the counter showing the oscillator frequency near 10 MHz, and the test oscillator itself.