Topics

2247A Counter/Timer Calibration question


tekscopegroup@...
 

Yesterday I turned on my 2247A and it immediately went into self cal and then reported a problem with the backup battery. I replaced that battery back in 2012 with a fresh one from BatterySpecialists.com, as the original one was already dead when I first got the scope. I thought its weird that the replacement barely lasted for 8 years. Guess they don't make them as they used to? I don't recall checking the date of manufacture of the first BR-2/3 replacement, or if it even had one at all, but will definitively check again when it comes out, and also record the date on the new replacement BR-2/3.

Anyway while I am going to go into the scope wanted to touch up a couple of things, like the step balance on the lower ranges, and the Counter/Timer calibration which I had noticed for some time now that it was consistently off by about 70Hz for a couple of years now, when compared to my rubidium oscillator, and which I know is dead on 10MHz as measured with another calibrated counter.

So according to the 2247A service manual (070-6367-00, page 5-15 / 147) it says to set the scope counter/timer to measure frequency, connect a TG501 set to 1uS time markers, which should be equivalent to a 1MHz signal. It then instructs to adjust C1904 (10MHz internal time base cal) "for a frequency readout on the scope between 999.994,10 and 999,995,10 KHz (0.999,994,10 to 0.999,995,10 MHz)".

My question is, why if I am using a signal of exactly 1MHz, do I have to calibrate the time base slightly lower? I would think that one would want to calibrate to match an exact reading of 1,000,000 instead. Can anyone shed some light on this? BTW the procedure as usual instructs the calibration to be done after warming up the scope for at least 30 minutes, so drift would be at a minimum when doing the calibration, and so I would assume that calibrating for a slightly lower frequency readout would not be related to initial cold start drift compensation. Or is this perhaps done for some long term drift compensation?

So why not calibrate dead on to the input signal frequency?


 

I did exactly that with mine, adjusted it to my GPS standard applied to the input, and ignored that manual procedure. It has worked fine since, BUT be aware that crystal is hanging out in the breeze, and is of no special quality. The shift in temperature from case open to closed is a serious jolt to the crystal (possibly what they are trying to fix with that procedure by setting it LOW). I find the best compromise I wind up with is roughly a reading 17Hz high after the scope has run for a while. A case hole to allow the adjustment with the case closed would be much better, but a tcxo would really be an improvement in this case.

all the best,
walter


tekscopegroup@...
 

Hi Walter,

Thank you for the information, and that exposed crystal certainly sounds like a reasonable explanation for the low spec frequency adjustment called for in the cal procedure with the case removed. Definitively would concur Tek might have made that adjustment accessible from the outside, ideally with a hole near the external 10MHz input jack would have made a lot of sense. Also guess the design budget for this model back then did not allow for a TCXO / OCXO, as these where probably more expensive back then? Nowadays one could get off ebay a decent TCXO from the US for under $20. I really haven't looked at the circuit, but perhaps as a viable upgrade one could think about building in either type of stabilized crystal oscillator hooking it up into the external 10MHz input, or if possible right into the circuit to retain the external input functionality.

Matter of fact just remembered that I do have two 10MHz TCXO's in my junk drawers, taken from old scrapped GPS receiver boards. One is labeled "RAKON TX0408R 10.000MHz", but has no external trim adjustment. Also have one this is physically larger and does have an adjustment. This last one is marked "GH-1711 10000.00KHz TOKYO DENPA CO. LTD" (now part of Murata). Could unfortunately not find any data on either of these, anyone familiar with them?

In any case to Tek's credit I have to say that as-is the Counter/Timer functions have been quite stable and very useful for the time I have owned this scope, except of course for that very small 70Hz long term drift which after 30 odd years would be perfectly acceptable as an age related issue. I like this scope very much and use it quite a bit more than others, including my 2465B, due mainly to all the versatile counter/time/voltmeter/cursor functions.

Thanks again.
-Alex


tekscopegroup@...
 

Since its so easy to open up this scope today removed battery without even having to move the board, and it only measures 0.11V. I don't understand how a backup battery can get this low in only 8 years. What looks to be the printed date code on the battery reads "2122" , not even sure what that is meant to be for a battery purchased in 2012. Last time I used the scope without issues must have been not 3 weeks ago.

Anyway, installed a temporary solution by using the battery that was taken out last year from the "My $25 In Poor Shape 3468A has arrived" DMM posting on the HP IO Groups forum. Battery still reads 3.043V open circuit (3.011 in circuit) so might as well use it while the fresh one arrives. Was not prepared to take a chance with this original battery guarding the cal constants of this DMM, but for this scope where it only really backups the last used configuration might as well use it for now, specially since the cal routine that kicks in on power up seems to put quite a burden on the screen while doing all the auto-cal routines with the intensity in what looks to be maxed out intensity.

On the other hand was looking at the position of the adjustment trimmer for the 10MHz calibration and it is squarely facing up into the top side of the cover. It would have been I think trivial to just add this one hole to be able to access this adjustment without having to remove the cover. I am still considering if I might drill a hole or not.


 

You should measure the current.

Regards

On 9/13/2020 8:18 PM, tekscopegroup@... wrote:
Since its so easy to open up this scope today removed battery without even having to move the board, and it only measures 0.11V. I don't understand how a backup battery can get this low in only 8 years. What looks to be the printed date code on the battery reads "2122" , not even sure what that is meant to be for a battery purchased in 2012. Last time I used the scope without issues must have been not 3 weeks ago.

Anyway, installed a temporary solution by using the battery that was taken out last year from the "My $25 In Poor Shape 3468A has arrived" DMM posting on the HP IO Groups forum. Battery still reads 3.043V open circuit (3.011 in circuit) so might as well use it while the fresh one arrives. Was not prepared to take a chance with this original battery guarding the cal constants of this DMM, but for this scope where it only really backups the last used configuration might as well use it for now, specially since the cal routine that kicks in on power up seems to put quite a burden on the screen while doing all the auto-cal routines with the intensity in what looks to be maxed out intensity.

On the other hand was looking at the position of the adjustment trimmer for the 10MHz calibration and it is squarely facing up into the top side of the cover. It would have been I think trivial to just add this one hole to be able to access this adjustment without having to remove the cover. I am still considering if I might drill a hole or not.

--
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus


tekscopegroup@...
 

On Sun, Sep 13, 2020 at 06:50 PM, Tom Miller wrote:
You should measure the current.

Regards
Guess you're assuming that the battery was nominal and that there is some form of above normal battery current leakage? Any idea what should be the expected current parameters? Would this even be possible given the circuit topology that is fed by the battery? Opinions welcomed. Thanks for the input.


 

I would expect the standby current to be much less than 10 uA and even much less than 1 uA depending on the part number of U2521. It should be a uPD4464C-20 . Better would be the L suffix. Check the ram data sheet.

Regards

Measure the current both with and without power applied.

On 9/13/2020 10:22 PM, tekscopegroup@... wrote:
On Sun, Sep 13, 2020 at 06:50 PM, Tom Miller wrote:
You should measure the current.

Regards
Guess you're assuming that the battery was nominal and that there is some form of above normal battery current leakage? Any idea what should be the expected current parameters? Would this even be possible given the circuit topology that is fed by the battery? Opinions welcomed. Thanks for the input.

--
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus


tekscopegroup@...
 

Agree, Ram stby current should be in that neighborhood if no other extraneous leakage. New battery is on the way, so will measure current before installing it.
BTW wanted to mention that in those 8 years since I have this scope, it was in storage without any use or power for about 3 years, could that have possibly sped up the premature demise of the first replacement battery?