Re: 2440 nvram (was: NVRAM back-up battery change on a 2715..... avoiding disasters)

Szabolcs Szigeti


Yes, i think we are on the same page now. I cas share some of my
The attenuator gain and the trigger adjustments only require a couple of DC
voltages. Given that the scope itself has something like 2% accuracy
specification, they do not have to be super accurate. Before I acquired a
scope calibrator, I simply used a precision power supply and measured the
voltage with a multimeter. The interesting part in this calibration is that
this is what provides traceability, as besides calibrating the
attennuators, the scope's internal 10V reference is compared to the
voltage, thus it provides a link to traceable voltage source. If there is
some (I think more than 2%) difference between the voltage provided during
the calibration and the scope internal reference, it will fail. For the
same reason the scope will indicate if any calibration was attempted, thus
traceability was broken.

The channel delay adjustment is about making sure that both channels have
the same delay. The scope is not very interested in having a very fast rise
signal, you can actually see that it simply tries to move the two channel's
waveform so that they are perfectly aligned with each other. I have
successfully performed this calibration with simple pulse generator of 5-6
ns riseftime. What actually is important here is to making sure that the
scope gets the same signal, so equal length cables should be used on both

The repet calibration is completely automatic no signal input is required
and it will succeed, unless there is some serous misalignment of the CCD

The CTE calibration is the most difficult, as the service manual refers to
PG506 settings to provide input. But it describes that it needs 100kHz and
6 div p-p. The scope sets itself up for this calibration step, which you
should not change, so if you don't have a PG506 you have to figure out the
input signal. I used a pulse generator to supply a 100kHz 200mV p-p signal
with some offset to provide the necessary signal. ( I hope I remember these
numbers correctly) You can see it on the screen so it is easy to find the
proper setting and the scope will complain if the signal is not good. My
experience is that it needs a clean rising edge, without much overshoot or
ringing. But this step can also be done without a PG506.

So the upshot of all this is that if you have at least a DC power supply
and a half-decent pulse or function generator it is possible to do the
ext-cal on this scope. A PG506 is useful, but it mostly makes the
calibration quicker, by requiring less fiddling with setting up the
calibration steps. Otherwise,it can perfectly be done without it. And you
certainly don't need any additional less common equipment, like TD ulser,
or time-mark generator.


Raymond Domp Frank <@Raymond> ezt írta (időpont: 2019. nov. 10.,
V, 1:03):

On Sat, Nov 9, 2019 at 11:16 PM, Szabolcs Szigeti wrote:

Hi Szabolcs,
You wrote:

The service
manual says that self cal + ext cal will bring the scope back to in spec.
It turns out that's correct.
Note: Ext Cal stores "Internal Calibration Constants" in (some sort of?)
NVRAM, which is exactly what I am talking about.
Ext Cal is no big issue but you need some (calibrated) equipment to do it.


Chapter 5 paragraphs 4.e and 4.f (page 5-19 in my SM) "Match Ch1 and Ch2
bw at remaining V/div settings". Sine Generator and 5x and 10x attenuators

Ext Cal (starting at page 5-23):
- "1. Attenuator Gain Adjustments". Needs calibrated voltage generator and
stores constants.
- "2. Channel Delay Adjustments". Needs fast rise calibration generator,
stores constants. I'd be surprised if a slow edge would be ok
- "3. Trigger Adjustments". Calibration generator needed. Stores constants.
- "4. Ramp (REPET) Calibration. Stores constants.
- "5. CTE Calibration. Needs Fast Rise Calibration Generator. Stores

From the above, I'd say that (Ext) Cal is not only a matter of pressing
some sort of "Self Cal" button(s). It involves connecting and using
calibration equipment and pressing buttons. As said, no big deal (a PG506,
possibly with TD pulser, will do as a generator and voltage source) but a
large part of calibration *is* lost if NVRAM is exchanged. Whether that's
serious is a matter of interpretation but I'd be surprised if an average
2440 will be "in spec" without re-establishing the NVRAM constants. Maybe
deviation is 5% max, as I've seen with analog 24xx'es, don't know.


Join to automatically receive all group messages.