Re: Adjustment of the power reading on the HP 8920A

#935

I would tend to trust the 8920 more than the Bird.

IIRC correctly the Bird wattmeters are +-5% of full scale ...not sure what slug you have in the Bird .? If that's a 100 W VHF slug you are +- 5 Watts? which pulls those numbers pretty close.

As was mentioned I would test/compare? the 8920 with some other device.? ?Usually an easy check is? a known good signal gen from @ 0, 10 and 20 dbm.? ?Make sure you zero the 8920 power meter first and set the scale to dbm as well.

Normally I have a number of power meters? and the 8920's I fix are very close.

You may have an issue on the RF i/O assembly ...not really an adjust per se on the watt meter.

I have had to replace leaky diodes in the power detect circuit? ...they actually met spec at higher levels but below 0 dbm the leakage became a factor.

Most of your typical through line wattmeters? are not supper accurate ....and then you have your cabling and terminator that needs to be verified as well.

I have? at least two of everything? so that if anything looks off I have another one to validate with?

Re: Adjustment of the power reading on the HP 8920A

#934

Hi Bill,

I listen to the normal gaggle of 2m repeaters so you can find me there. Tied up with WARG (again). Re your calibration against the Bird, is the Bird a good reference??

Happy to have a chew of the rag :)

73 de?
VK6XRE
Robyn

Re: Adjustment of the power reading on the HP 8920A

#933
Edited

Your Bird 43 is likely far less accurate than your HP. The elements include a potentiometer that is prone to shifting or going ‘rusty’. Some even fail open circuit. I took an element apart once and was shocked at how highly geared the 10 cent pot was - from memory needing about 20k of 100k full range.

Find a properly calibrated power meter before you touch your 8920. Mine was within a percent or two after over a decade without cal.

If anything, adjust your Bird. But then you have two challenges: fiddling with the element, or the meter, as they both have adjusters.

73, Colin G4CWH

Show quoted text

On 8 Jan 2024, at 07:11, William Macham via groups.io <westwill1@...> wrote:

?Hello all,

I have finally got my instrument going well, NO FAILURES  :)

The one thing left is to adjust the reading to agree with the value of watts at the rf input, measured with my Bird 43 watt meter.

I presume the adjustment is in the "Configure" menu ... ?

Can someone advise me as to the procedure to make the correction? 27.6 watts in @ 144 Mhz, reads about 21 on the screen.

Regards

Bill M.     VK6IC

Re: 8924C POWER SUPPLY DEAD

#932

Hi Robyn,

I see we are both VK6 ... and proud owners of these great instruments. I have been into the PSU?on my 8920A also.? At last all is going sweet.?

?Perhaps we could support each other with some local dialogue.

?I am QTHR and a member of NCRG.

Regards

Bill M.

?

On 29/12/2023 7:24 am, Robyn Edwards wrote:

Show quoted text

Hi all,

?

My 8924C died as well. Thankfully the excellent PDF and a $0.40 47 uF, 50 v high temp electrolytic fixed it.? What I will say is that you can remove and replace that cap in situ if you are gentle and patient.? An electric screw driver is very useful.

?

73

de VK6XRE

Adjustment of the power reading on the HP 8920A

#931

Hello all,

I have finally got my instrument going well, NO FAILURES? :)

The one thing left is to adjust the reading to agree with the value of watts at the rf input, measured with my Bird 43 watt meter.

I presume the adjustment is in the "Configure" menu ... ?

Can someone advise me as to the procedure to make the correction? 27.6 watts in @ 144 Mhz, reads about 21 on the screen.

Regards

Bill M. ??? VK6IC

8921A PCBs required

#930

Thanks, Yuri

I suppose this 28C256 originally attached in the unit is broken.

?

James

8921A PCBs required

#929
Edited

Hi John,

Thank you.

?

Hmmm, I gotta say that I'm suspecting the processor board right now.

?

Yes, if the blank 28C256 is attached, checksum error only pops out for the very first time. After that first boot, the checksum error is gone.

?

Without recovery of CAL data, I think swapping processor board might be worthless.

?

But I'll give it a try. I gonna buy a processor board to eliminate the last possibility. There are parts mules in the on-line market.

?

Thank you again.

?

James

Re: 8921A PCBs required

#928

开云体育

Hello,

Friday, December 29, 2023, 10:31:48, chinesejames1990 wrote:

Thanks. Umm should I fiddle with my programmer to see if I can make some adjustable supply voltage on the chip?
,_._,_

There is a lot of development programmers that have adjustable program and verify voltages (last one I used had it in programming method customization setup). But if programmer shows pin connection errors, first get a multimeter with a low voltage diode test mode and look if you can "see" ESD protection diodes between GND and VCC to each pin (maybe except Vpp). I had one plastic "windowless" one time programmable EPROM with a faulty pin wire welding, and it was successfully read in a cold state with a piece of ice in a small plastic bag on the top of chip.

It worth trying (carefully) to extract data from this EEPROM, because otherwise whole unit became a set of spare boards, and most of them without cal data.

--
Best regards,
Yuri ? ? ? ? ? ? ? ? ? ? ? ? ?mailto:yuri@...

Re: 8924C POWER SUPPLY DEAD

#927

When my 8924C died a few years ago, it was the little supply that keeps the OCXO warmed up. The rest of it wouldn't work without that. Getting to that little supply was a headache, but a couple of caps fixed it.

The last time it died has been a total nightmare. Swapping boards from my 8924E only made both of them bad. A second replacement board got it working, sort of. It still says it has bad cal data, but seems to work otherwise. The power supply of the 8924E blew up. I still need to address both issues.

Show quoted text

On Sun, Dec 31, 2023 at 3:34?AM Robyn Edwards <vk6xre@...> wrote:

Hi all,

?

My 8924C died as well. Thankfully the excellent PDF and a $0.40 47 uF, 50 v high temp electrolytic fixed it.? What I will say is that you can remove and replace that cap in situ if you are gentle and patient.? An electric screw driver is very useful.

?

73

de VK6XRE

Re: 8924C POWER SUPPLY DEAD

#926

Hi all,

?

My 8924C died as well. Thankfully the excellent PDF and a $0.40 47 uF, 50 v high temp electrolytic fixed it.? What I will say is that you can remove and replace that cap in situ if you are gentle and patient.? An electric screw driver is very useful.

?

73

de VK6XRE

Re: 8921A PCBs required

#925

James,

It sounds like you may have a bad processor board ...not sure if that is what also damaged the EEPROM.? Still could be another issues ....second on my list is the memory board? which on the 8920A has the firmware and the NVRAM ( battery backed up )

Firmware should throw? a checksum error if it was bad or not reading correctly ...you already tried that and didn't have those errors anyway.

If you put in a blank write enable EEPROM you will also see numerous checksum errors on the initial start for the calibration data but on the second boot that should be gone since default data was loaded in.? check in your programmer options that the EEPROM is software write enabled ....sometimes that is a setting you have to enable.

If it was the EEPROM causing the initial error that should have fixed it ...the diagnostics will fail with the default calibration? data.

At this point if you were in the US I would suggest sending it in for repair.

The initial problem can be fixed and the calibration may be recoverable but it sounds doubtful at this point.

Blown calibration data may turn the unit into parts since the effort to recover factory tuning may be more trouble than the unit is worth.

Still has value as parts ...just not as much as if the calibration was good.

8921A PCBs required

#924

Hi Yuri,

?

Thanks. Umm should I fiddle with my programmer to see if I can make some adjustable supply voltage on the chip?

?

Thanks,

James

8921A PCBs required

#923

Thanks John,

Sorry for the late reply.

Yes, I gave it another shot these days.

?

First, I put an empty new 28C256 on and turned on the unit; errors remained the same.

?

Second, I used another programmer to read the original 28C256 and it couldn't even read it. The programmer popped errors about "bad lead connection" on some specific pins. The programmer is surely good.

?

Well, maybe this is it. Got nothing more I can do. Maybe the unit is scrapped.

?

Thanks again.

James

Re: 8924C POWER SUPPLY DEAD

#922

I can add a "me too".? My power supply died in 2009 and I paid someone to fix it, it just died again and this time I decided to tackle it myself this time.

So far I've confirmed that was the problem by providing +15V from an external bench supply as suggested in the PDF.? That made it possible to turn on and all of the self tests passed!? (Whew !)

Two possibly new pieces of information for anyone running into this problem:

1. After disconnecting the bench supply and turning the power "off" I still measure a perfect 15.00 volts on standby and it turned back on successfully.
2. My green LED still lights up even with the bad +15v (I measured 4.9V), but it's a LOT brighter at 15V.? So if your LED is still lit do not assume that the cap on the standby board is not the problem.

73's Skip WB6YMH

Re: 8921A PCBs required

#921

开云体育

Dear Yuri,

??? You could send us the 28C256 IC and we could evaluate it at no charge. It's possible the IC is corrupt. We can read it and transfer the data to another 28C256 and see if that works. You could do that yourself with a prom programmer. You would need the 28C256 - we have many. What are the symptoms of the problem? Rick

On 12/25/2023 3:44 AM, Yuri Ostry wrote:

Show quoted text

Hello,

Monday, December 18, 2023, 16:43:43, chinesejames1990 wrote:

My guess is CAL data is broken due to faulty original EEPROM 28C256. And the worst thing is, I can't read it out in good shape nor can I re-calibrate the unit. Various selftest errors may be caused by bad CAL data.

Is there any way to restore it, or any way to measure in order to re-calibrate the whole unit just like what factory did for people like us-individuals?
,_._,_

If you have access to a programmer with valiable supply voltage, try to read your chip many times while decreasing supply voltage gradually, and write every read image separately. I had similar problem with Xicor X2816 chip in Marconi 2955, and finally I was able to generate correct image by analyzing files read from such "leaking" chip. If you are really lucky, you may find voltage at which chip will read almost stable with correct data (say 9 of 10 reads will be the same). It is same method that we used to recover "leaky" EPROMs in the past - IIRC, I already wrote here or in time-nuts list with more details.

Regards,
Yuri.

--
Best regards,
Yuri ? ? ? ? ? ? ? ? ? ? ? ? ?mailto:yuri@...

IBasic programming

#920

Happy Holidays to all group members

I am a proud new owner of an HP8920A. My main use will be for various amateur radio applications. I have been reading as much documentation as I can find on the internet. I am amazed by the capabilities of this instrument. The IBasic compiler offers some intriguing and creative possibilities. Has anyone developed useful IBasic programs for the 8920A?

Re: 8921A PCBs required

#919

开云体育

Hello,

Monday, December 18, 2023, 16:43:43, chinesejames1990 wrote:

My guess is CAL data is broken due to faulty original EEPROM 28C256. And the worst thing is, I can't read it out in good shape nor can I re-calibrate the unit. Various selftest errors may be caused by bad CAL data.

Is there any way to restore it, or any way to measure in order to re-calibrate the whole unit just like what factory did for people like us-individuals?
,_._,_

If you have access to a programmer with valiable supply voltage, try to read your chip many times while decreasing supply voltage gradually, and write every read image separately. I had similar problem with Xicor X2816 chip in Marconi 2955, and finally I was able to generate correct image by analyzing files read from such "leaking" chip. If you are really lucky, you may find voltage at which chip will read almost stable with correct data (say 9 of 10 reads will be the same). It is same method that we used to recover "leaky" EPROMs in the past - IIRC, I already wrote here or in time-nuts list with more details.

Regards,
Yuri.

--
Best regards,
Yuri ? ? ? ? ? ? ? ? ? ? ? ? ?mailto:yuri@...

Re: 8921A PCBs required

#918

replacing the calibration data is going to be tough for the whole unit.? It will load defaults when it hits a checksum issue ...every section in the cal data has a checksum it's so important.? ?However these default values will often cause the unit to fail the RF diags in critical areas.? In a lot of areas the defaults are close enough but there are tuned circuits that need the modules specific data .

On some of the later cards the cal data was on the board ...see the service manual.

May sure the EEPROM is SW write enabled ...there is a feature on those chips to disable the write besides the dip switch on the processor board.

I have never seen an EEPROM fail completely ...usually only specific addresses will fail to verify.

I would double check if you can with another programmer.? You really need that data which is why I mention backing it up.

I have two very good IC programmer units that I use to cross check on issues like this.

Make sure the IC is correctly aligned going back in as well .....on one of the processor boards ( A vs. B ) it seems like it goes in backwards.

People always ask why I have so much test gear? and I usually go back to the old military saying " one is none, two is one and three is enough "

Like I said you could put in a blank EEPROM as long as it is write enabled and the unit will load defaults? ...that should get you past the initial self test failure but the unit will probably fail the RF diags.

Re: 8921A PCBs required

#917

Hi all,

First, I'd like to thank all of you who gave me ideas and clues.

So here is what I've done these days:

1. I ran the selftest program contained in the ROM, and I almost every error message popped out.
2. I checked the PSU supply rail and all the voltages were within criteria.
3. I think I found the smoking gun, that EEPROM 28C256 containing CAL data is broken:
? ? a) I used my programmer to read out the CAL data various times, but each time I read it the data file is different. (Using my BIN file comparator software, and the programmer (checked and it's in good working condition)
? ? b) I pressed and hold MEAS REST + Hz keys and power up the unit, so that all status could be reset as factory without affecting CAL data and ROM. I tried several times as well, but each time the screen displays this error message: "CAL DATA CHECKSUM FAILED; Default DATA is applied". Ummm not sure the exactly wording, but something like that.? ?(That screen is very hard to capture, it just blinks and is gone).
? ? c) I wrote all the BIN files read out from original 28C256 to new 28C256s one by one, and put the new 28C256 into the unit one by one, but got no luck; problem remains.
4. Bought 8 27E040 PROMS and rewrote the firmware using the BIN files downloaded from the FILES SECTION. Everything works good except the fault remains exactly the same.

My guess is CAL data is broken due to faulty original EEPROM 28C256. And the worst thing is, I can't read it out in good shape nor can I re-calibrate the unit. Various selftest errors may be caused by bad CAL data.

Is there any way to restore it, or any way to measure in order to re-calibrate the whole unit just like what factory did for people like us-individuals?

Thanks in any advance!

Regards,
James

HP8920a PLL 1GHz out of lock; it's a problem of the divider by 50

#916

Hi ,

I have a spectrum analyzer HP8920a and during the self test RF DIAG I have the message "VCO 1Ghz Not locked". I checked at the board A15 and I have made some measurements. The output of the divider by 50 (signal 10_PLL10 pin 11 U10 and pin2 U27) is not working; there is no signal. At the entry of the divider by 50 the 500MHz?signal is correct.

The problem is the schematic in the user manual is different from the board. In the user manual the divider uses U27 MC1201 and U18 MC1038, and on the board there are U27 74AC191 and U18 "1820 8762".

I try to find some datasheet about this circuit 1820 8762 I don't find anything on the web. For the 74AC191 no problem; I have the datasheet.

To test the board I just put the power supply voltages +12V, -12V, +5V on the P2. In that manner it's possible to easily?test the board A15.

So my PLL is out of? lock because it's missing the signal "10_PLL10" pin 11 of U10. Has anyone ever encountered this problem? If someone has the datasheet of U18 "1820 8762" and the schematic of this divider by 50, it would be nice.

Thank you for your help.

F4EYM.