Optoelectronics 3000A Product Review - September 1993
Copyright (c) 93,94 Brian Mork -- "not to make money, but to keep it that 
way for others."

>>> HISTORY
June 19th of 1993, I distributed a product review of Optoelectron-
ics' 3000A handicounter.  It had bad dBm calculations.  A 29 July 
distribution fixed these errors and the text was cleaned up in other 
ways.  More complete input sensitivities were reported.  This version 
includes subsequent perspective after use of OE's M1 handicounter (see 
the M1 product review of this same date -- it is a pared down version of 
the 3000A). Damien Thorn (Internet drdamien@delphi.com) wrote a non-tech-
nical review article on several of OE's products (covering more why you'd 
want them and what to do with them once you own them) that appeared in 
the August 1993 issue of Nuts & Volts.

>>> ABSTRACT
A 10Hz-3GHz feature-laden counter/timer/frequency meter could be a good 
addition to your shack if you don't have the functionality yet. There are 
some disallowed configurations that give you less than you may think if 
you casually read the laundry list of features.  Input specifications 
need more attention or at least acknowledgement of real-life problems.

>>> CONSTRUCTION
The case is an extruded U-channel of aluminum with a flat top plate and 
end plates.  The end plates screw into the U-channel with four screws, 
one of which is a machine screw.  This one is by the A input, apparently 
designed to provide a better "bite" into an interior copper shielding 
plate around the A input.

A single 9.2cm x 12.6cm circuit board lies under the faceplate, directly 
mounting the three buttons and six slide switches, which, in turn, poke 
through the front plate.  A piggyback (piggyfront?) board holds the LCD 
display.  There are two main chips--the OE10 vaulted in Optoelectronics' 
marketing literature and an SC87C51.  My 8751 is labeled M1.3000A 5.20.93 
VER 2.1 with a little sticky label.  The two main chips and a transformer 
on the board have a white chalky substance around them as if they were 
hand soldered and the flux vapors weren't cleaned off.  Hope it doesn't 
corrode.  Probably 200 additional components, mostly surface mounted, 
also occupy the circuit board.

The six batteries are stuck to the back shell (double sided tape?).  The 
two battery wires terminate on the circuit board with a separable connec-
tor.  The batteries' weight and the solid aluminum case give the unit a 
solid feel.  Neither the wall-wart charging unit or the 3000A itself have 
a charging light.  For the extra resistor and LED, I would like one.  Too 
many times I've had some master switch, wall switch or loose plug inad-
vertently negate my effort to charge batteries.  And when do you find 
out?  When you needed the unit to be charged.

There are three plastic sheets stuck on the external top and face of the 
unit.  My top piece was peeling off slightly.  The front bottom one was 
crooked enough to let Button 2 not bounce back up.  Slight trimming fixed 
that.  I can't tell if the white lettering is applied on top of the plas-
tic or behind the plastic.  I hope it's behind so that they don't wear 
off.  All the aluminum is painted black --not drippy, but it has the ap-
pearance of thick enamel.  Personally, I would much rather have black 
anodized so there would be no possibility of chipping paint.

>>> OWNER'S MANUAL
It's five sheets of photocopied 8-1/2" x 11", printed in landscape mode 
and folded in half.  It's short, but operation of the 3000A is easy, so 
maybe a lot isn't needed.  I would propose inclusion of a few diagrams 
and tables rather than all linear text description.  Included in the man-
ual are:

     Features description............... 2 pages
     Specifications..................... 1 page
     Operation.......................... 2 pages
     Data Interface..................... 2 pages
     Block schematic.................... 2 pages
     Operation hints.................... 4 pages
     Battery & Power.................... 1 page
     Calibration........................ 1 page
     Warrantee.......................... 1 page
     Service & Return................... 1 page

A loose sheet of paper describes how to turn on the backlight option.

I wanted to know what I could plug into this meter.  Nowhere in the specs 
or owner's manual are the max input ratings. After talking with the fac-
tory (toll call is small compared to the purchase price, but still is 
annoying), I find the 1 Mohm inputs can handle a maximum on the order of 
50 volts AC+DC.  The Model 3000 (not the 3000A) specified 100V RMS as the 
maximum.  Anyhow, I feel good about plugging in TTL.

DON'T plug TTL into the inputs if the 50 ohm switch is selected!  More 
subtly, be sure you *unplug* it before you slide the switch to the 50 ohm 
position.  The same phone call revealed that the 50 ohm inputs handle a 
maximum of +15 dBm. That's 1.2 volts RMS in a 50-ohm system, correspond-
ing to 3.4 volts Peak-to-Peak if the waveform is a sine wave.  Another 
user was told by a factory technician that "a 5W HT connected to the 
counter input for a short time would not damage the counter."  Five watts 
is about +37 dBm!  Does the amount of time matter?  The correct numbers 
need to be in an Owner's Manual, especially since the warranty excludes 
"damage to the input circuitry from the application of an excessive input 
signal."

I would suggest making no direct electrical connections to the 50 ohm 
amps.  Only antenna type inputs should be used when the 50 ohm amps are 
selected.  I've wrapped an 8" wire (snake like) around cordless phone 
antennas, and that works ok, but even this would be *way* to much for a 
100W HF transmitter.  Be careful.  Full scale bar graph ranges from 2 - 4 
mV, depending on the frequency.  If you stay on scale, you're safe, but 
most of the valid input window is above "full scale."  In the range 4 mV 
to approximately 1200 mV, you'll be increasing power "in the blind."
 
>>> OPERATION
Some non-obvious combinations of functions and inputs is the order of the 
day.  This is kind of weird, but is intentional to give you the best sen-
sitivity possible, using a number of specialized input amps.

Turn it on, holding down a Button 2 if you want the optional backlight to 
come on.  Button 3 selects the gate time (resolution).  With a slide 
switch, you select either 1 Mohm (Hi-Z) or 50 ohm (Lo-Z) amplifiers.

If you select Hi-Z inputs, Button 1 selects either/both A and B inputs.  
Button 2 rotates between FREQUENCY, PERIOD, INTERVAL, and RATIO.  Input A 
measures only 10Hz - 50MHz, with sensitivity of a few tens of millivolts.  
Input B measures only 10Hz - 10MHz, with a sensitivity of better than ten 
millivolts.

If you select Lo-Z inputs, a signal strength bar-graph indication auto-
matically appears and you must select one of three input amps:

    Input A only, from 10MHz - 220MHz (0.3 to 5mV sensitivity)
    Input A only, from 10MHz - 880MHz (divide by 4 prescaler)
    Input B only, from 500MHz - 3000MHz (new amp, div by 16 prescaler)

On one of OE's product literature flyers, they speak of a 1-200MHz input.  
A phone call confirmed that this range exists only as a typo on the pro-
duct literature flyer.

In parallel with all the above input selections, two switches let you 
operate in either of four modes.  I'll call them modes 00, 01, 10, and 
11, representing whether FILTER and CAPTURE, respectively, are off (0) or 
on (1).

   00 (FILTER off, CAPTURE off)  Operates as described above.  This is 
      the single mode of the original Model 3000.  Meter shows updated 
      values even if it's counting noise.  Gate light blinks continuous-
      ly.

   10 (FILTER on, CAPTURE off)  PERIOD, INTERVAL and RATIO are disabled.  
      Computer logic looks for meaningful oscillations, providing a new 
      frequency value only if a legitimate count is acquired.  A higher 
      level of filtering can be selected by holding down Button 3 during 
      power up.

   11 (FILTER on, CAPTURE on)  Same as mode 10 except nothing happens 
      until you arm the system by pressing Button 2.  Once you do this, 
      the word "frequency" flashes on the display until a legitimate 
      count is snagged.  A three-memory buffer (the display plus two 
      more) keeps previous values.  They are referred to as X, A, and B.  
      X contains the most recent, A the second most recent, and B the 
      third most recent. When a fourth value is snagged, the value in B 
      is lost.

   01 (FILTER off, CAPTURE on)  In this mode, no measurements are taken.  
      Button 2 rotates between the three memories recorded in mode 11.

>>> PRODUCT SUPPORT
In some sense, there's not much to support.  The meter either works or it 
doesn't.  Operation is very straight forward.  I don't call to harass 
service personnel and I *do* read the manuals first.  I did call to find 
out about the input characteristics as discussed above and was, well, not 
impressed.

Female answers and I ask for service or tech help.  After 2 minutes, the 
line drops dead.  I call back.  Same female passes me on to Ray quickly 
this time.  I ask Ray about input specs.  He says that, yes indeed, that 
has been inadvertently left out of the manual.  Stand by.  Three minutes.  
Ray comes back with info.  The 1 Mohm inputs are "plus 50 volt AC plus 
DC", the 50 ohm inputs are "plus 15 dBm."  Yes, that's what I wanted!  
Follow up:  What do you mean by 50v AC+DC?"  Hmmm.  He didn't know; 
that's what the engineer told him.  Ok.  Is it capacitively coupled?  
Uhh.. the inputs go right into a "MAR6 amplifier."  Could you please send 
me a copy of just the front end prior to the MAR6 amps?  Pause... "Well, 
there's a capacitor that goes to ground."  Yes, that's the kind of info I 
want.  What size?  Maybe a pf or two?  "It's a 50v electrolytic."  Eeee  
wrong person.  Would he send me just a copy of the front end schematic?  
"No, we don't have anything like that I could send you."  Ok, well thanks 
anyhow.  Bye.

In retrospect, I should have asked for the engineer who was sourcing this 
information.  Inputs are listed in the specifications as 1MHz, 30pf.  
This seems believable since it mimics a normal oscilloscope input.  But 
are they AC or DC coupled? With what size, type & rating capacitor?  I 
was hoping to use the INTERVAL or RATIO mode with Input B held open so I 
can count events on Input A.  At Dayton a year back, I spoke with a man 
named Bill Owens who seemed to have a good handle on the company's histo-
ry and the capability of the product line.  I wish his knowledge was con-
veyed better to potential buyers.

>>> REAL LIFE
I have two regimes of interest for which I want this counter.  The first 
is modem and ultrasonic work in the KHz to 100 KHz range.  The second is 
radio VFO work, requiring MHz up to GHz.  I characterized the 3000A with 
a 20Hz - 1MHz function generator, my Yaesu 757GX/Vectronics VC300DLP com-
bo, and a Hewlett-Packard VHF/UHF generator.

--- low frequency ---
For the less than 1MHz work, the Hi-Z inputs are used.  I have a 1MHz 
TXCO standard in my shack that has an output floating on the back side of 
a little toroid transformer.  It gives out approximately +-0.5v into my 
500 Kohm load (3000A and oscilloscope in parallel).  My 1MHz output shows 
up as a stable 6.00 Mhz on the 3000A.  What?!  Yes.  Regardless of the 
filter selection (none, medium, or full), six megahertz kept being dis-
played.  The waveform was not symmetric, I'll give that much, but it was 
very stable and clean from other high frequency hash.  It looked sort of 
like this:

       |\      |\        The waveform stayed the same regardless of  
 0 V___| --    | --      whether the 3000A was hooked up.  The 3000A 
       |   |   |   |     counted the same regardless of whether the  
      -.   .---.         o'scope was hooked up. P-P voltage was 1.5v.

The 1MHz output from my TXCO was a square wave ranging between 0 and 2 
volts.  Regardless of the 3000A filter selection, it showed a frequency 
of 3.0 MHz.  Another user working on a homebuilt oscillator also reports 
extreme sensitivity to harmonics.  This harmonic sensitivity, combined 
with extreme sensitivity makes the instrument virtually useless when us-
ing any frequency multiples in your project:  you won't know which signal 
is being locked onto.

I fired up a variable frequency / variable amplitude function generator 
and o'scope combination.  The generator went to the scope with a BNC T-
tap half way that plugged into the 3000A.

Using the A input, I needed incredulously large amplitudes--WAY above the 
specified <20 mV.  Using the B input was even worse.  The procedure used 
was to start at about 100 mV and go down first.  If a stable count held, 
I reduced the PP voltage until spuratic readings were obtained.  If 
100 mV was giving spuratic or known-to-be-wrong numbers, I turned the 
amplitude up until a stable reading was obtained.  The resultant measured 
PP voltages (asterisks indicate I never was able to get a good count):

     Frequency      Input A             Input B
     ---------      -----------         -------
      20 Hz           2 V (0.78 VRMS)      *
     200 Hz           1 V                  *
     350 Hz           0.6 V                *
      60 KHz          50 mV                4 V
     500 KHz         <10 mV                1 V

Square wave counting was similar.  For instance, 2 V PP was needed at 
10 KHz to get a good count with Input A.  In all cases, the scope showed 
clean signals with no high frequency hash.  Testing at another ham's 
shack who had a fancy HP generator and scope yielded the same results 
within 10%.  Stable operation of the Hi-Z input is mandatory to use the 
interval and period functions of the 3000A, and this condition does NOT 
seem to be met.  I consider the non-frequency modes to be historical car-
ry-overs from the product lines' 7226 counter heritage rather than useful 
features of OE's OE10 custom chip.

I own a PK232 radio modem and decided to see if the meter was good enough 
to tune the audio tones.  With full output from the PK232 (about 0.5v PP 
on the scope, giving 140mV RMS on a Fluke DVM good to 10KHz), I stuck the 
tones into Input A and Input B with Hi-Z amps selected.  The 2200Hz tone 
showed as 2265+-5Hz (A) and 23KHz (B), respectively.  The 1200Hz tone 
showed as 1820+-15 (A) and 31KHz (B), respectively.  No, I did not adjust 
the PK232.

--- high frequency ---
My RF test goal was simple.  How accurate is my 757GX VFO?  Working near 
10MHz, the Lo-Z A input (only one suitable for this measurement) has an 
advertised sensitivity of <5mV.  With a 12" dangling wire near the coax 
out the back of the radio, the counter was dominated with a signal at 
approximately 104 MHz (103.9 FM?).  Later tests 25 miles across town 
again showed that with a wire pigtail antenna, it was being dominated 
with this station.  I bought an RS15-577B FM trap and put it in line.  
Now a 68 MHz interference dominates, probably the local Channel 4 TV 
station.

I went back to the Hi-Z inputs.  At this point I was trying to determine 
the electrical length of some coax.   A 1/4 wavelength of coax is termi-
nated with a short circuit, and fed in parallel with a 50 ohm load with a 
mid level RF power (see page 74 of the June 93 QST; the goal is to get a 
1:1 SWR).  The high current node at the terminator is the point I was 
trying to pickup RF with the 3000A.  Twenty turns of perfboard wire 
around a pencil stepped up the voltage enough for the meter to register 
stable counts when the pickup coil was positioned near the end of the 
coax shorting wire loop.

Characterization of the three 50-ohm amps with an HP generator showed 
solid frequency locks at these random check points:

       Fast Amp              Middle Amp             Slow Amp
----------------------  ----------------------  ----------------------
 800MHz, -47dBm, 1.0mV  100MHz, -65dBm, 0.12mV  150MHz, -60dBm, 0.22mV
1000MHz, -43dBm, 1.6mV  450MHz, -54dBm, 0.45mV

Attempts to do PERIOD, INTERVAL, and RATIO measurements were unsuccessful 
due to the skittish behavior of the Hi-Z inputs.

>>> OPTIONS
You can buy a precision (0.2 ppm vs 1.0 ppm) timebase for an extra $100.
What used to be a $45 backlight option now seems to be a non-optional 
extra expense.  September advertisements still quote the lower price for 
the standard model, but the sales personnel won't sell one that way.

--- time base ---
There are two reasons why I did not want to buy the precision time-base
option.  One is because a RATIO mode is provided and I already have a 
precision 10MHz TTL oscillator scavenged from an old LORAN navigation 
board.  10MHz also happens to be the maximum frequency Input B accepted 
in the ratio mode (all ratios are A/B).  What a coincidence.  Works for 
me!

The second reason is simply that the 1ppm timebase is specified to age at 
a rate of 1ppm/year.  A 0.2 ppm option would be degraded to the 1.0ppm 
option in only about 10 weeks.  In any case, I usually don't need to know 
the last 2 Hz on a 10000000 Hz signal.

--- backlight ---
When it's on, I can hear a switching circuit bringing up a high voltage 
for the luminescent display.  It turns off after 10 seconds of inactivity 
and comes on again when some button is pressed or frequency is acquired.

It looks good.  Worth $45? The native display is plenty readable and the 
extra light actually makes reading it harder from angles "above" the 
counter.

>>> GOTCHAS
The serial interface is unidirectional.  You send it a CR and it sends 
back 10 digits and a decimal point, in ASCII, 2400bps, 8 bits, no parity, 
1 stop bit.  It only works in FREQ mode.  It provides only the most re-
cent number, with no indication of whether this is another sample or the 
same number it just sent you after your last request.  The interface can 
sink 1.6mA and source 0.06 mA.

Excellent amplifier sensitivity isn't everything.  Sometimes it's even a 
burden.  For decent counting, the signal you're monitoring must exceed 
the noise (combination of *all* other RF signals in the bandpass of the 
selected amp) floor by 10-15dB.  Specified sensitivity ranges from -57dBm 
to -11dBm.  The input amp is limited to +15dBm.  Ambient noise, including 
FM stations hovers about -11dBm.  Play with those numbers and you'll see 
that the window for a good count can be pretty small. Try to measure a 
cordless phone near your transmitter or computer monitor?  Probably not.  
If you live near a broadcast station, good luck.  I don't consider these 
numbers a design flaw given the intentionally wide frequency range of the 
amps, but it definitely affects your day-to-day operation.

In Damien's review, he emphasizes the importance of a limited bandwidth 
antenna.  Take this recommendation seriously.  Wide range frequency cov-
erage means susceptibility to noise.  A bandwidth limiting antenna helps 
mitigate this problem.

>>> SUMMARY
The at-your-door price for a 3.5" x 5" circuit board seems a bit high, 
but the counter is specified to do what I need it to do, plus a few 
options.  Resolving the skittish Hi-Z input problems is a must.  The M1 
gets around this by simply not having multifunction capability.  Paying 
the extra $100 for a 3000A with specified capabilities that in real life 
are marginally usable deserves a second thought. 

All models are available only direct from the manufacturer in Florida. 
Contact Optoelectronics: 5821 NE 14th Avenue, Ft Lauderdale, FL 33334.  
800-327-5912 or 305-771-2050.  FAX 305-771-2052.  Makes you want to 
dial ..2051 and see who you get, doesn't it? :) 

73, Brian Mork (Opus-OVH)     ARO KA9SNF@ka7fvv.#ewa.wa.usa
                              Internet BMORK@opus-ovh.spk.wa.us
                              6006-B Eaker, Fairchild, WA 99011
