Re: 1A5 differential comparator 'sponginess'

Dave Wise
 

Chuck, 1A5 Vc is only 5V. You might be thinking of Type Z, 100V.

Chris,

I played with my 1A5 in a 545 mainframe and saw what I think you saw.
By the way, the manual uses either a 545 or a 545A, because the graticule is only +/-3 divisions.

I set it to its native sensitivity of 5mV, which has a +/-5V common mode range.
(2mV and 1mV are achieved by a bandwidth-sacrificing second-stage gain increase.)

1. Vc has a time constant. It's due to C114. This will show up anytime you move the POLARITY switch.

2. As long as you can keep all of the trace on screen, Type 1A5 handles its full rated common-mode voltage range with full-speed recovery and no shift.

3. Driving the trace far offscreen does cause a DC shift.
I would not be surprised if, in my case at least, part of it was in the mainframe. You go far enough, something will cut off or saturate.
In any case, DC operating points change, causing some parts to heat up and others to cool. With 5mV sensitivity, you'll notice it. The designers did some work to minimize it; what we see is what they couldn't compensate without blowing the budget.

4. Read the manual. See Table 1-1 Electrical Characteristics, page 1-2.
DC Shift Due to Overdrive <= 1% of +/- overdrive signal or 10mV whichever is smaller
Overdrive Recovery Time Within 10mV of DC shifted level after 1us. [Ed. Note: Return to "DC shifted level" not "original level".]

5. There's a procedure for checking it. See Section 5 Performance Check, Step 14, page 5-5, Check Overdrive Recovery Time and DC Shift Due to Overdrive.
A-B with B grounded.
Prepare a clean 10kHz square wave whose positive level is 0V and negative level is -5V.
Set the instrument to 10mV. (5V is 500 divisions.) Wait one minute with no signal, center the trace, then apply signal to A.
Trailing level of trace should be within 10mV of graticule center within 1s after applying the signal.
Then leading edge over/undershoot should be within 10mV of center within 1us.
Repeat for B.

Dave Wise

-----Original Message-----
From: TekScopes@groups.io [mailto:TekScopes@groups.io] On Behalf Of Chuck Harris
Sent: Thursday, September 26, 2019 4:56 PM
To: TekScopes@groups.io
Subject: Re: [TekScopes] 1A5 differential comparator 'sponginess'

I committed an editing faux pas. The Note about the AC
position should be moved down a few paragraphs as shown
below.

-Chuck Harris

Chuck Harris wrote:
I fear that you may not understand the purpose of the 1A5.

The 1A5 is a differential amplifier plugin.

Its reason for being is to allow you to measure a small signal
that is riding on top of a larger common mode signal.

A common mode signal is a signal that both the A input, and
the B input have in common, and typically, you would like to
ignore.

An example is if you took a small battery powered oscillator,
that outputs 1Vp-p, and connected its signal return output to
the 120V power line. So now the 1Vp-p signal is riding on top
of a 120V sine wave. In an ordinary scope plugin, you would not
be able to see the 1Vp-p signal for the huge 120V sine wave.

XXXXXXXX (Note, this is what the AC position is for on most plugins...)

However, if you used the 1A5, you could connect the A input to
the battery powered oscillator's signal output, and the B input to
the battery powered oscillator's signal return (which is also
connected to 120VAC), and put the 1A5 in A-B mode, and be able to
see the battery powered oscillator's signal full screen, without
any sign that the 120Vp-p "Common Mode" signal even exists.

The 1A5 has 3 basic modes of display:

A-Comparison_Voltage
A-B
B-Comparison_Voltage

The A-B mode is the major mode of this plugin. The two comparison
modes are additional modes that tektronix achieved by the addition
of a highly adjustable low current HV supply, called the "Comparison
Voltage"

The comparison voltage takes the place of the input not mentioned on
the Display switch. So, for A-Comp, the voltage is connected between
ground and the B input. For B-Comp, the voltage is connected between
ground and the A input.

This allows you to create your own common mode voltage to cancel out
large DC offsets. Something that is commonly needed when working with
tube circuits, and switching power supplies.
<<< HERE >>> (Note, this is what the AC position is for on most plugins...)

Recovery from input overload conditions is one of the most serious,
highly ignored, problems with all amplifiers. The 1A5 is no exception,
and is not imbued with any great recovery abilities... as you have
discovered.

The 1A5 is NOT able to work with signals much larger than the scope's
screen.

-Chuck Harris

cmjones01 wrote:
I've just become the proud owner of a 1A5 differential comparator
plugin for my 535A scope. It's a wonderful thing and its only real
fault was that the variable volts/cm pot was made of that awful
crumbly Tek plastic and had fallen apart. With that temporarily
shorted out while undergoing repair, the rest works really well after
a dose of DeoxIT D5.

However, it has one odd behaviour which I can't explain. The whole
point of the 1A5 is being able to work with signals much larger than
the scope's screen. However, if I feed this one a large positive
signal, the trace drifts upwards somewhat over a period several
seconds, then drifts back down slowly after that signal is removed.
The same happens the other way with a large negative signal. It
happens on both A and B inputs, and also when applying/removing a
large comparison voltage.

For example, I have it running right now set to 0.1V/cm. With a signal
of anything up to about 1V square wave from the calibrator, the
'ground' level of the trace stays constant, subject to the usual
long-term drift I'd expect. However, if I apply a 10V square wave,
still with the plugin set to 0.1V/cm, the 'ground' level wanders
slowly up about 0.8 of a division. It takes several seconds to do
this. If I turn the calibrator back down to 0.5V, the trace wanders
slowly back down. The signal is well within the plugin's rated
common-mode range.

Can anyone explain this? Are they all like that? My 7A13s, which
presumably have a similar architecture, don't suffer from it.

I'm wondering if there's some leakage somewhere which is charging
something up slowly, when given a large offset input signal.

Chris

Join TekScopes@groups.io to automatically receive all group messages.