OpenBCM V1.06 (Linux)

Packet Radio Mailbox


[C.D.G. OpenBCM]

 Login: GUEST

G8MNY  > TECH     14.10.18 10:45l 244 Lines 12245 Bytes #4 (0) @ WW
BID : 41767_GB7CIP
Subj: Scope & DMM Calibrator
Sent: 181014/0830Z @:GB7CIP.#32.GBR.EURO #:41767 [Caterham Surrey GBR]

By G8MNY                                      (Updated May 17)
(8 Bit ASCII graphics use code page 437 or 850, Terminal Font)

This device lets you test the basic accuracy, calibrate Y1 gains, & frequency
response of a scope's input attenuator, & allow accurate adjustment of the
compensation trimmers, & also test Meter calibration.

The design gives accurate DC or 1kHz square wave voltages from 10V p-p down to
1mV p-p in 10, 5, 2, 1 steps.

The rise time (not fall time!) is better than 100nS @ 10V, but the attenuator
Rs & stray capacitance can double this on the 5 & 2V ranges.

It can also be used for checking other DVMs or other high input Z meters.

         2 Stage  납  Level  납1kHz Astable척 Mode  납 11 Volt   12-14V
Output<캑 4 Decade 척Calibrate척 Cal Freq   척Control척Reference쳐< DC
        쿌ttenuator냅컴컴컴컴켸 Cal Square 냅컴컴컴켸읕컴컴컴컴
        읕컴컴컴컴켸           읕컴컴컴컴컴켸

Based on a simple PNP astable, reference voltage source, & output attenuator..

  10V    |   5k        
   5V o   |  o컵컴훓                           OFF
        S|   3k         10V                      o
   2V o   |  o컵컴훒   1k CAL              DC  MODE
        C|   1k      Preset          旼컴컫컴컴컴훟 SWITCH
   1V o   |  o컵컴훐                __  __       <컴컴컴컴   e컴 +1218V
        R|   500               D2 \_/  \_/ D1  o          \ /      @ 5mA
500mV o   |  o컵컴훏                            SQUARE   컴쩡
        E|   300        BC214                 WAVE        TR1
200mV o   |  o컵컴훎      旼컴컴컴컴(컴컴좔쩡컴컨컴쩡컴      PNP
        E|   100        e\쿟R3         /e               
100mV o   |  o컨컴쩡a    PNP쳐커     쳐컴캑PNP              
    100k N|              ./        TR2\               1M
 50mV o   |  o훓       쳐켸                              
         |      100  __                                
 20mV o  P|  o훒      \_/      쳐컴컴)컴늘컴   100u+       
         |         D3   4n7         4n7       ===       
 10mV o  L|  o훐       쳐캑쳐컴)컴컴캑                     
  5mV o  A|  o훏            120k   120k                   
         |                                       __+   
  2mV o  T|  o훎          Set 읕47k켸               /_\'   
         |           10k  Square/\       4k7          Z1  
  1mV o  E|  o훍          pre Pot                    11V 
     憫  |                      /\ Freq                 
    1k Ct |                    ,20k Preset                      훩e
    ATTENUATORS                ASTABLE                 REF    CURRENT LIMIT

TR1 (BC214) provides a limited current source (Approx 5mA) controlled by the
1M base resistor, to put current into an 11V Zener Z1 & the rest of the
circuit. Although this current source varies a little with the supply &
temperature it is much better than just a resistor, & hence the zener reference
voltage (5+6V zener) gives say 10.8V 10mV, is kept fairly accurate!

If the +10V range is not wanted the circuit can be worked with a 5.6V zener, a
PP3 battery & suitable attenuator changes! For higher voltages different
circuits are needed with diodes to protect the bases from over reverse voltage.

When the DC mode is selected D1 (1N4148) powers the circuit, while D2 inhibits
TR2 from conducting & hence T3 must conduct in the normal way providing an
identical +ve output voltage to that when the circuit oscillates. D1 voltage
drop should be of no consequence due to the constant current from TR1.

The TR2 & TR3 (BC214) make up the normal if up side down astable circuit. Diode
D3 isolates the charging up of T2 base capacitor from the T3 output, so keeping
the edge turn on very fast.

The frequency is determined by both sets of C & R (4n7 & 120k) time constants
so the Cs should be polyester type for best thermal stability. Any in-balance
of the square wave (not 1:1 ratio on a scope or 2nd harmonic nul on AF spectrum
analyser) can be adjusted by trimming the set square pot. Using a common 20k
preset to trim the joint bias will allow the frequency to be trimmed to exactly
1kHz with a frequency counter.

This is done using an accurate DVM ( >1M Input Z). Set the mode switch to DC &
set the output to switch to 10V, adjust the 1k Cal pot for 10.000V. Check that
this DC is the same value as the 1kHz square wave rises to, on an oscilloscope.

The attenuator resistance values are made up so that 10mV/1k ohm gives easy
numbers to work with. But the 5k=10k//10k, 3k=3.3k//33k or 2x 1.5k etc. need 2
Rs to make the exact value from the E12 resistor series. All the Rs are
soldered around the switch. I used a screened 2 bank 13 step switch, other
switches will do with the Rs made up the values to suit.

The 6 positions out of the 1st 7 positions on the 1st wafer bank are paralleled
up to the last 6 positions, which sees the 2nd wafer bank switch in a 100:1
attenuator, which uses high enough values not to significantly load the 1st
attenuator values. Open wire low capacity wiring is needed for all the wiring
in this area, e.g. NO "neatly tied up wiring forms", just short looped wires!

For accurate work (1% or better) the R values can be trimmed with much higher
value Rs to ground, or 10V using an accurate DVM as for the initial 10V

If the 5V output impedance is too high (2.5k=5k/5k) for some loads, put a 5k6
from the 5V point to ground & change the 10k//10k for 2k7//68k, the 68k value
needs to be selected on test to give 5V once the 10V CAL is setup again.

For a good square wave on the low signals end of the attenuator, a screening
plate is required between the dual bank switch, & possibly a trimming "Ct" to
ground (1-20pF) across the 1k to over come crosstalk from the 10V to 1mV

This assumes you are OK with working inside live scopes with HIGH VOLTAGES!

Connect the calibrator with a very short coax cable. With the scope at maximum
Y1 gain setting, but without any added gain multiplier (as these reduce the
bandwidth), select a suitable calibrator square wave signal level to give a
25cm display, using timebase controls to show 0.2cm/mS to give a large steady

STEP 1. Check front panel Y1 Gain is set to "Cal" position. Adjust internal Y1
preset gain (not high gain X10) for correctly calibration height avoid display
parallax error. Adjust preamp C/Rs tweaks for best (ideal) square rising edge
response. Check the X10 (or whatever) gain option is also accurate adjust that
high gain preset, if there are separate C/Rs for X10 also adjust.

N.B. some scopes have HF C/Rs that will not be adjustable with a 1kHz square
wave but need faster 1MHz or RF frequency sweeps to set up.

     TO LITTLE HF             CORRECT               TO MUCH HF
     ____         _       ______        __      `\.___
  ,/'         ,/'       |      |      |        |      |      |    | = faint
  |      |     |         |      |      |        |      |      |    | Verticals
  |      |     |         |      |      |        |      |      |
  |      `\.___         |      |______|        |      |   ___|

STEP 2. Switch Y input attenuator to next position & up the calibrator level to
suit, check deflection calibration (do not adjust scope calibration unless you
have suitable replacement scope attenuator Rs!), adjust the correct Y
attenuator's series C (Ctrim) for best (ideal) square rising edge response.
There may be two Cs, a series one (Ctrim) & a parallel one (Cstray), this one
seems to do nothing at this stage!

Repeat this for the first set (order) of input attenuators.

                   Typical x10 x100 x1000    Typical x2 x3 or x5
SCOPE                    Attenuator              Attenuator
INPUT  o)컫캑쳐컫컴>o컫컴컴캑쳐컴컴쩡훟<컴>o컫컴컴캑쳐컴컴쩡훟<컴 AMP INPUT
 BNC       AC    SW      Ctrim     SW   SW     Ctrim     SW     Input Z
Input Z   읕o \켸      쳐훀series컴캑          읕훀series컴캑         1M //
1M //      DC                                                      30pF
 30pF                 ===Cstray   Rshunt                  Rshunt
        Both Attenuators banks are straight through wires for x1.

STEP3. Higher order attenuators (x10, x100, x1000) must "see the same load
capacitance" of the first order attenuators so they can only be adjusted once
the sensitive ranges have been done.

STEP4. Equalising the scope input capacitance (Cstray) across these ranges is
best done with a x10 scope probe. Set the Y attenuator gain attenuator to its
max sensitivity, with the x10 scope probe, view a suitable sized square wave.
Adjust the probe's own trimmer for the best waveform. Now step up the scope Y
attenuator over the ranges with suitable input levels, if the square wave shape
changes, adjust the unused (did nothing before) parallel Y caps (Cstray) for
the same shape. Remove the scope probe.

Now repeat step 3 & 4 until there is no tweaking needed. Go & adjust Y2!

Make sure the Timebase velocity control is in its "Cal" position. Find the LF
timing preset & adjust (sometimes it is the X gain) so that the 1kHz waveform
fits a 1mS/Division range (~_ cycle in each square). Check on some of the
other ranges for a mean calibration. For HF timebase calibration you will need
an RF signal generator.

If it has these modes, then they can be calibrated for the correct gain to,
in the same way as STEP 1, but U just see 2 dots unless U can put a locked
timebase into Y.

Using a square wave to calibrate RMS meters is not as complex as it may seem.
On AC, most meters are MEAN reading, but calibrated in RMS for sine wave use.
The sine wave mean is 0.636 of the peak, compared to the RMS value of 0.707 of
the peak, so the readings are scaled higher. The error is exactly 1.11 which is
called the "form factor".

  100%_     _      __PEAK         100%_ __________  __RMS = 5V
 70.7%_   /' `\     __RMS         +5V  납          
 63.6%-           --MEAN             납             1:1
                                     납            SQUARE
       납                              납          
   0% -납- - - - -- - - - -       0% -납 - - - - - - - - - -
                    \._./          -5V_ COUPLED   __________

So back to the square wave of 10V p-p (DC component removed with a series cap
if needed) this is a 5V peak square wave, which actually has the same RMS value
as 5V DC, & will read 5V on a true RMS meter. But a MEAN reading meter which is
scaled x1.11 higher & will read 5.55V AC.

The same goes for a square wave, 10V peak 50% of the time (what the square wave
should be!) should read 5V, but 7V RMS. So you can determine if a meter is MEAN
reading DC or not.

 100%_ __________            __PEAK = 10V
      납             1:1     --RMS = 7.07V
      납            SQUARE
 50% -납 - - - - - - - - - - --MEAN 5V
         50% ON   50% OFF  
  0% _ COUPLED   __________  __0V

With the Calibrator on DC mode, any reasonably high impedance meter can be
tested for basic accuracy & damaged or burnt out meter calibration Rs detected.

Also see my Tech buls on "Oscilloscopes", "Scope Probes" & "Kelvin Varley
Voltage Divider".

Why Don't U send an interesting Bul?

73 De G8MNY @ GB7CIP

Read previous mail | Read next mail

 21.10.2018 02:04:44lGo Back Go up