Re: HP 141T Power Supply Help Needed
Diode to chassis is how I was measuring it.??? I do see expected voltage at the cathode side of my makeshift rectifier.? 17.5v.
The 24v last night was the fuse..?? here's the most recent checks.
Q9
E 1.5
B 1.27
C .62
Q10
E 0
B .6
C 2.2
Q11
E 1.66
B -2.3
C 2.2
Bill.
For the sake of clarity, what exact points in the circuit are you
measuring between?
Across each diode (anode to cathode) or between one side of each
diode vs Chassis Ground?
If the latter, you have a big problem.? The diode Cathodes, "should"
be some 6 (or 9) Volts Positive wrt chassis.
What is shown during a Diode Test on a DMM, depends on the DMM, and
what else is still connected to the diode in question.? Such tests
usually are only valid on a diode "out of circuit" (or with one end
disconnected at least.)? Even then, a faulty diode (that breaks down
/ leaks above some voltage well under it's spec') will often seem OK
on a DMM.
Regards.
Dave B.
From: Bill
Berzinskas
Date: Sat, 14 Sep 2024 04:38:53 GMT
Took both side panels off, dug the transistor holder out a bit
more and proceeded to trace the wiring.? ?It all checks out, but I
now have strong reservations about the status of CR1 and CR2.? ?I
believe I am seeing?
Anode - 24v dc / 17v ac
Cathode - 0.5v DC
Assuming it should register something on my DMM, since its a diode
and all..? ?Nothing.? ??
I take it these are the source of all my woes?? ? Not sure what
the heck I was measuring before..?
--
Created on and sent from a Unix like PC running and using open source software:
|
The usual problem - degraded O rings!
They seem around 1mm ID 0.5mm CS (so 2mm OD) or maybe .039" ID, .024"
Cross-section, which can readily be found in Butyl-rubber as watch
crown gaskets, but there are various comments that they need to be
Viton/FKM in order to last, but so far I've not found a supplier
willing to sell me a sensible number (in UK).
Does anyone know of one or will watch crown gaskets do?
Alan
|
Re: HP 141T Power Supply Help Needed
Bill.
For the sake of clarity, what exact points in the circuit are you
measuring between?
Across each diode (anode to cathode) or between one side of each
diode vs Chassis Ground?
If the latter, you have a big problem.? The diode Cathodes, "should"
be some 6 (or 9) Volts Positive wrt chassis.
What is shown during a Diode Test on a DMM, depends on the DMM, and
what else is still connected to the diode in question.? Such tests
usually are only valid on a diode "out of circuit" (or with one end
disconnected at least.)? Even then, a faulty diode (that breaks down
/ leaks above some voltage well under it's spec') will often seem OK
on a DMM.
Regards.
Dave B.
From: Bill
Berzinskas
Date: Sat, 14 Sep 2024 04:38:53 GMT
Took both side panels off, dug the transistor holder out a bit
more and proceeded to trace the wiring.? ?It all checks out, but I
now have strong reservations about the status of CR1 and CR2.? ?I
believe I am seeing?
Anode - 24v dc / 17v ac
Cathode - 0.5v DC
Assuming it should register something on my DMM, since its a diode
and all..? ?Nothing.? ??
I take it these are the source of all my woes?? ? Not sure what
the heck I was measuring before..?
--
Created on and sent from a Unix like PC running and using open source software:
|
Re: HP 141T Power Supply Help Needed
Re:-
"That discrepancy between A2Q9.C and Q4.C is certainly a typo, or perhaps the two points were measured at different times and the mains voltage changed, and no-one noticed that they're by definition the same."
Under different load conditions perhaps?? (Plug-ins present, or not.)
I see there is also mention of 2V ripple on the +ve unregulated side (Q4.Q9 collector.)? Presumably when under some unspecified load.
(I don't have a copy of the full manual.)
Regards.
Dave B.
--
Created on and sent from a Unix like PC running and using open source software:
|
N9010A 32-bit XSA Software for Windows 7
Hey All,
?
I'm just wondering if anyone may have a copy saved somewhere of this software.
?
I'm trying to bring an old N9010A with a 32-bit x86 processor onto a later OS than Windows XP (Windows 7) but it seems that Keysight either never made a Win7 x86 version of the software or has let it fall off their website and the internet.
?
The N9010A will run Windows 7 but the Installers for the older software are gated to check the Windows Version so I can't just re-run the WinXP installer and when you move from Windows XP to Windows Vista (as part of the OS upgrade process) it breaks the app install (it expects to run the installer again which then errors out).
?
The right way to address this is to take the N9010A from WinXP to Win7 and then reinstall the app but all the versions hosted on the Keysight website expect that you have an x64 process board in the N9010A.
?
Any help is appreciated.
?
Thanks,
?
TonyG
|
Re: HP 141T Power Supply Help Needed
For S&Gs, I jumpered in a set of SB520 and that brought the -12.6v output up to ~23v!? ? I went ahead and ordered 1N3209s.?
-23v is a bit concerning, but I will check the pass trans one more time and remeasure everything in the morning.? ?
On Sat, Sep 14, 2024 at 12:38?AM Bill Berzinskas via <bberzinskas= [email protected]> wrote: Took both side panels off, dug the transistor holder out a bit more and proceeded to trace the wiring.? ?It all checks out, but I now have strong reservations about the status of CR1 and CR2.? ?I believe I am seeing?
Anode - 24v dc / 17v ac
Cathode - 0.5v DC
Assuming it should register something on my DMM, since its a diode and all..? ?Nothing.? ??
I take it these are the source of all my woes?? ? Not sure what the heck I was measuring before..??
?
?
|
Re: HP 141T Power Supply Help Needed
Took both side panels off, dug the transistor holder out a bit more and proceeded to trace the wiring.? ?It all checks out, but I now have strong reservations about the status of CR1 and CR2.? ?I believe I am seeing?
Anode - 24v dc / 17v ac
Cathode - 0.5v DC
Assuming it should register something on my DMM, since its a diode and all..? ?Nothing.? ??
I take it these are the source of all my woes?? ? Not sure what the heck I was measuring before..??
?
?
|
? Hi Aldri, ? ? For your second converter, when you do debugging it's easy to make a mistake, there is so much manipulation. You have to keep the positive side, this part is set and you have a second spare converter. ? For the YTF, it's not easy, centering the spheres is not enough, don't forget that it's a bandpass filter. So you have to align the filter for a certain reference frequency, but which one, we could assume that it's in the center, but not necessarily, it could be at the beginning for example (lowest frequency). The filter will deform according to the frequency, it's about finding a happy medium. The YTF calibration will find the coefficients of a polynomial trying to have the best for all frequencies. The level differences will be compensated during the frequency flatness calibration. However, don't forget that if you have too much loss to compensate, the DANL test could become out of spec. ? The fact that you use a precise external power supply is not a problem for the alignment of your YTF. The important parameter of this power supply is its stability. ? I had sent you a picture of my 8595E, here it is again, it represents a cumulative test of the frequencies: yellow curve -> test at 3GHz cyan curve -> test at 5GHz magenta curve -> test at 6.5GHz We can see that on the left, there could have been an improvement by re-adjusting one of the spheres, but nothing is certain. The insertion loss varied from 2 to 3 dB between the minimum and the maximum of these 3 frequencies. You have similar results, I think your YTF has good results...nothing is perfect :) ? You say that you have problems in the high band, what are the frequencies in play. from 2.9GHz to 6.5GHz? (highband). For the second converter, the path loss for the highband is -1.8dB. After this point two bands go through the same path, so if the YTF is correct, either there is a problem with the YTF control source, the first converter or the second converter (highband section). You should be able to easily find where the excess loss is. ? You are at the end of your troubles, don't give up, you will find it. ? ? --------------------? Here is the verification of this second converter after reassembly. The flatness is good, only 0.03dB for 5MHz width at output IF 321.4MHz (lowBand). Multiplier x6 for 3600 MHz, adjusted for best match (lower loss at 321.4MHz), very sensitive adjustments. Path for 321.4MHz highBand was verified too. ? One sweep with 50 MHz span, another with 2950MHz span¡to see the stop band. The loss for IF highband include loss of filter and RF switch, around -1.8dB. ? Yves ? Yves, thanks for the info about the 2nd converter? After having installed the one bought on the bay?I checked?the original one?on the bench and it?seems to work normally I'm starting to suspect the blindness on the LOW BAND was due to a bad connection I did when the repair was suspended years ago ¡ ouch ! Meanwhile I'm investigating?a certain deafness on the HIGH BAND (from 2.4 to 11 dB) I measured the insertion loss of the SYTF I had realigned and found figures?from 5.2??to? 8.5 dB on the range? 3 to 6 GHz These figures?seem not too bad?and?the AMP CAL routine should be able to recover them ...unless there is some hardware problem in the generation of the tuning current (I performed the insertion loss measurements by manually tuning? with an external precision power supply)
toggle quoted message
Show quoted text
----- Messaggio originale ----- Inviato: 13/09/2024 07:51:28 Oggetto: Re: [HP-Agilent-Keysight-equipment] HP 5086-7906 YTO
Hi Adri, ? A friend of mine had a faulty 8595E, the problem comes from the option card 130. But I took the opportunity to look at the second converter, so here are some details on this module. ? Regards Yves ? ? hello Yves,
I received the 2nd converter and, after testing it on the bench (if you are interested I will give you the two test configurations) I installed it and voil¨¤, the machine is back to working like in the old days
It passes the amplitude and frequency calibrations and the confidence test without any problem
As it happened before the death of the YTO the analyzer still has a certain "deafness" after 5.5GHz
By virtue of the alignment of the spheres that I did a few weeks ago The defect is milder than it was then and I think it is due to a sub-optimal loss of insertion of the SYTF A3A2 As proof of the "deafness", while all the calibration routines, including the CAL YTF routine are performed without problems, the CAL MXR is not completed and gives the message CAL: HARMONICS INSUFFICIENT In fact I attach a screen of the 300MHz calibration signal from which it would seem that the very last harmonics are missing
Thank you in advance if can you send the analogous screen of your analyzer
In a few days I will receive a more reliable 6GHz RF generator than the current one and I will use it to try to perfect the alignment
Yves, I think that now you can reassemble and close your analyzer :-) Many thanks for all the measurements that you sent me and that were useful to trace the problem!
Adri ----- Messaggio originale ----- Inviato: 08/09/2024 01:35:49 Oggetto: Re: [HP-Agilent-Keysight-equipment] HP 5086-7906 YTO
Hi Adri ? Here are the measurements connected to the first 2 converters as a reference. ? For the second converter I get a voltage of -1.678V instead of -3.5V as written in the service manual, I would say that we have similar voltages for this particular point. It seems far from the nominal point, but there may be an error in the manual. ? You are close to the goal, and this module may be repairable if there is no micro circuit, I have never opened a second converter in this series, but it is a possibility. ? When you open your second converter, send me pictures, I'm curious. ? Yves ? dear Yves, I didn't realize that you had made the measurements expressly to help me! I owe you then a double (even triple) thanks! Especially considering the fact that the front end of this instrument is really very dense, complicated and delicate! After my last email, in addition to doing the (essentially mechanical) intervention on the SYTF, I continued to check the input signal through the various sections by alternately injecting 300MHz @-20dBm and 5GHz @0 dBm The signals were found to be in specification up to the A3A9 bandpass (where the two paths join to form the intermediate frequency of 321.4MHz). At this point I could see the 5GHz signal (converted), present at that point at -47dBm (the specification says -43/-56dBm).
Instead, no trace of the 300MHz input signal
Investigatine?the second converter A3A4 in more depth
I checked?the signals at its input
On J1 there is 3.9214GHz at -46dBm? --->OK
On J3 there is 321.4MHz @ -37.7dBm? --->OK
On J4 I did not check 600MHz @-1dBm but the fact that the converter works in band 1 tells me that the oscillator is present and is in specification
The power supply is quite OK (5V are actually 4.75; 10V are 9.98 and the DC that controls the internal switch that should be +15V? (band 0) and -3.5V?(band 1) is actually +14.60V and -1.35V
This last value is rather far from the nominal and would have worried me BUT band 1 is exactly the one that works Conclusion: it would seem that the culprit is the second converter and in particular (probably) the internal switch that has the sole task of letting the signal? (which is already at 321.4MHz and does not need to be converted)?pass from its input to the output
In the next few days I will test the converter that meanwhile?has been?disassembled?; In the meantime I found a cheap one on the bay that should arrive in about ten days
Yves, I have a question: on the block diagram A3A4 is represented as a block that has 5 ports of which one (J5) would seem to be the connector that carries the power and the DC that controls the switch and that connects to A7 Analog Interface through the W35 cable
Actually,?this connector on the real component, does not have a name
Instead,?the name J5?, stamped on the case , indicates an SMB connector that has the writing LO TEST and in my analyzer?it is the one that?conveys the 600MHz LO signal? Now,?on the block diagram the connector that conveys the 600MHz signal has the name J4 (in my analyzer, the J4 connector was disconnected) This confusion of names and roles made me doubt that years ago, when I had disassembled this section I had not reassembled it correctly thanks again and best regards ----- Messaggio originale ----- Inviato: 06/09/2024 23:21:52 Oggetto: Re: [HP-Agilent-Keysight-equipment] HP 5086-7906 YTO
Hi Adri, ? I had sent you the measurements of the RF input-output mixer signals verification, I had disassembled the High Frequency section of my 8595E to be able to provide you with this information. If you have no more questions relating to this section, I will close my 8595E and do a flatness check to see if everything is correct. ? Yves ? ? thank you very much for all the suggestions and information with which I will continue my investigation
In the meantime I received the 2.5mm socket wrench and after?a little mod to adapt it?for the purpose (photo), I proceeded to align the spheres of the SYTF
I must say that the operation done under the microscope is quite simple
The nuts must be loosened (not too much) and with a small screwdriver that acts as a delicate lever, ?the supports of?the?spheres can be moved so that the coupling ring completely shadows them (i.e. the sphere is exactly in the center)
Removing the gold mesh was not necessary
The insertion loss has improved enormously:
at 5 GHz it is -6dBm at 6GHz it is -10dB (before the signal starting from this frequency was no longer visible) at 6.5 GHz it is only -3dB
I will install it when I have solved the problem of "blindness" in band 0 since the calibration constants that the instrument has in memory (in practice the tuning current data) are relative to the original SYTF and the "new" SYTW requires a? SYTF CAL is done
|
Hi Razvan,
thanks also for the service note. I have an E4407B, I had not
noticed that it was the same second converter.
Yves
toggle quoted message
Show quoted text
-----Message d'origine----- De?: [email protected]< [email protected]> De la part de Razvan Popescu via groups.io ·¡²Ô±¹´Ç²â¨¦?: 13 septembre 2024 03:49 ??: [email protected]Objet?: Re: [HP-Agilent-Keysight-equipment] HP 5086-7906 YTO Hello Yves, This second converter assembly is also used in the 8563E and also ESA E4407B spectrum analyzers. In fact I think it is common for all 6.5/13.5/26.5GHz versions. Mine had failed in the E4407B and it looks like it is related to an incorrectly manufactured transformer inside the assembly. I attached the Agilent Service Note related to this issue. Part number affected is 5086-7958. Thanks for the nice PDF with all explanations. Regards, Razvan On 13/09/2024 07:51, Yves Tardif via groups.io wrote: Hi Adri,
A friend of mine had a faulty 8595E, the problem comes from the option
card 130.
But I took the opportunity to look at the second converter, so here
are some details on this module.
Regards
Yves
|
Re: HP 141T Power Supply Help Needed
Comments are based on a scan of 00141-90914, 141T mainframe S/P 1615A, printed June 1976, file downloaded from BAMA.
The positive side of C6 varies with mains voltage.? If the regulator is working, it's whatever it takes to make the negative end be -12.6V .
There are complications when you try to compare winding AC voltage with DC output.? Capacitance, ESR, winding resistance, diode characteristics, load.? You can calculate a rough approximation only.
That discrepancy between A2Q9.C and Q4.C is certainly a typo, or perhaps the two points were measured at different times and the mains voltage changed, and no-one noticed that they're by definition the same.
Dave Wise
toggle quoted message
Show quoted text
Sounds good to me.
?
Mike
?
?
The schematic for the HP 141T shows all voltages referenced to ground.? If that is the case, you know the voltage at the positive side of C6.? It is listed as 5.9VDC.? So you can infer that the voltage across C6 is 18.6V.? From there,
you can find what the AC voltage across the winding should be, correct???
Also, in the schematic they show collector of Q9 at 6.61VDC and collector of Q4 at 5.9VDC.? But apparently Q4 and Q9 are tied together at the collector by wire 968.? What is the reason for the 0.6V drop?? Or is this an error???
|
Yves, thanks for the info about the 2nd converter? After having installed the one bought on the bay?I checked?the original one?on the bench and it?seems to work normally
?
I'm starting to suspect the blindness on the LOW BAND was due to a bad connection I did when the repair was suspended years ago ¡ ouch !
?
Meanwhile I'm investigating?a certain deafness on the HIGH BAND (from 2.4 to 11 dB)
?
I measured the insertion loss of the SYTF I had realigned and found figures?from 5.2??to? 8.5 dB on the range? 3 to 6 GHz
?
These figures?seem not too bad?and?the AMP CAL routine should be able to recover them ...unless there is some hardware problem in the generation of the tuning current (I performed the insertion loss measurements by manually tuning? with an external precision power supply)
Will see ...
Adri
?
?
?
?
toggle quoted message
Show quoted text
----- Messaggio originale -----
Inviato: 13/09/2024 07:51:28
Oggetto: Re: [HP-Agilent-Keysight-equipment] HP 5086-7906 YTO
Hi Adri,
?
A friend of mine had a faulty 8595E, the problem comes from the option card 130.
But I took the opportunity to look at the second converter, so here are some details on this module.
?
Regards
Yves
?
?
hello Yves,
I received the 2nd converter and, after testing it on the bench (if you are interested I will give you the two test configurations) I installed it and voil¨¤, the machine is back to working like in the old days
It passes the amplitude and frequency calibrations and the confidence test without any problem
As it happened before the death of the YTO the analyzer still has a certain "deafness" after 5.5GHz
By virtue of the alignment of the spheres that I did a few weeks ago The defect is milder than it was then and I think it is due to a sub-optimal loss of insertion of the SYTF A3A2
As proof of the "deafness", while all the calibration routines, including the CAL YTF routine are performed without problems, the CAL MXR is not completed and gives the message CAL: HARMONICS INSUFFICIENT In fact I attach a screen of the 300MHz calibration signal from which it would seem that the very last harmonics are missing
Thank you in advance if can you send the analogous screen of your analyzer
In a few days I will receive a more reliable 6GHz RF generator than the current one and I will use it to try to perfect the alignment
Yves, I think that now you can reassemble and close your analyzer :-) Many thanks for all the measurements that you sent me and that were useful to trace the problem!
Adri
----- Messaggio originale -----
Inviato: 08/09/2024 01:35:49
Oggetto: Re: [HP-Agilent-Keysight-equipment] HP 5086-7906 YTO
Hi Adri
?
Here are the measurements connected to the first 2 converters as a reference.
?
For the second converter I get a voltage of -1.678V instead of -3.5V as written in the service manual, I would say that we have similar voltages for this particular point.
It seems far from the nominal point, but there may be an error in the manual.
?
You are close to the goal, and this module may be repairable if there is no micro circuit, I have never opened a second converter in this series, but it is a possibility.
?
When you open your second converter, send me pictures, I'm curious.
?
Yves
?
dear Yves, I didn't realize that you had made the measurements expressly to help me! I owe you then a double (even triple) thanks! Especially considering the fact that the front end of this instrument is really very dense, complicated and delicate!
After my last email, in addition to doing the (essentially mechanical) intervention on the SYTF, I continued to check the input signal through the various sections by alternately injecting 300MHz @-20dBm and 5GHz @0 dBm
The signals were found to be in specification up to the A3A9 bandpass (where the two paths join to form the intermediate frequency of 321.4MHz). At this point I could see the 5GHz signal (converted), present at that point at -47dBm (the specification says -43/-56dBm).
Instead, no trace of the 300MHz input signal
Investigatine?the second converter A3A4 in more depth
I checked?the signals at its input
On J1 there is 3.9214GHz at -46dBm? --->OK
On J3 there is 321.4MHz @ -37.7dBm? --->OK
On J4 I did not check 600MHz @-1dBm but the fact that the converter works in band 1 tells me that the oscillator is present and is in specification
The power supply is quite OK (5V are actually 4.75; 10V are 9.98 and the DC that controls the internal switch that should be +15V? (band 0) and -3.5V?(band 1) is actually +14.60V and -1.35V
This last value is rather far from the nominal and would have worried me BUT band 1 is exactly the one that works
Conclusion: it would seem that the culprit is the second converter and in particular (probably) the internal switch that has the sole task of letting the signal? (which is already at 321.4MHz and does not need to be converted)?pass from its input to the output
In the next few days I will test the converter that meanwhile?has been?disassembled?; In the meantime I found a cheap one on the bay that should arrive in about ten days
Yves, I have a question: on the block diagram A3A4 is represented as a block that has 5 ports of which one (J5) would seem to be the connector that carries the power and the DC that controls the switch and that connects to A7 Analog Interface through the W35 cable
Actually,?this connector on the real component, does not have a name
Instead,?the name J5?, stamped on the case , indicates an SMB connector that has the writing LO TEST and in my analyzer?it is the one that?conveys the 600MHz LO signal?
Now,?on the block diagram the connector that conveys the 600MHz signal has the name J4 (in my analyzer, the J4 connector was disconnected)
This confusion of names and roles made me doubt that years ago, when I had disassembled this section I had not reassembled it correctly
thanks again and best regards
----- Messaggio originale -----
Inviato: 06/09/2024 23:21:52
Oggetto: Re: [HP-Agilent-Keysight-equipment] HP 5086-7906 YTO
Hi Adri,
?
I had sent you the measurements of the RF input-output mixer signals verification, I had disassembled the High Frequency section of my 8595E to be able to provide you with this information.
If you have no more questions relating to this section, I will close my 8595E and do a flatness check to see if everything is correct.
?
Yves
?
?
thank you very much for all the suggestions and information with which I will continue my investigation
In the meantime I received the 2.5mm socket wrench and after?a little mod to adapt it?for the purpose (photo), I proceeded to align the spheres of the SYTF
I must say that the operation done under the microscope is quite simple
The nuts must be loosened (not too much) and with a small screwdriver that acts as a delicate lever, ?the supports of?the?spheres can be moved so that the coupling ring completely shadows them (i.e. the sphere is exactly in the center)
Removing the gold mesh was not necessary
The insertion loss has improved enormously:
at 5 GHz it is -6dBm at 6GHz it is -10dB (before the signal starting from this frequency was no longer visible) at 6.5 GHz it is only -3dB
I will install it when I have solved the problem of "blindness" in band 0 since the calibration constants that the instrument has in memory (in practice the tuning current data) are relative to the original SYTF and the "new" SYTW requires a? SYTF CAL is done
|
Re: HP 141T Power Supply Help Needed
toggle quoted message
Show quoted text
From: [email protected] < [email protected]> On Behalf Of Frank Mashockie Sent: Friday, September 13, 2024 3:01 PM To: [email protected]Subject: Re: [HP-Agilent-Keysight-equipment] HP 141T Power Supply Help Needed ? The schematic for the HP 141T shows all voltages referenced to ground.? If that is the case, you know the voltage at the positive side of C6.? It is listed as 5.9VDC.? So you can infer that the voltage across C6 is 18.6V.? From there, you can find what the AC voltage across the winding should be, correct??? Also, in the schematic they show collector of Q9 at 6.61VDC and collector of Q4 at 5.9VDC.? But apparently Q4 and Q9 are tied together at the collector by wire 968.? What is the reason for the 0.6V drop?? Or is this an error???
|
Re: HP 141T Power Supply Help Needed
Dave,
?
The schematic for the HP 141T shows all voltages referenced to ground.? If that is the case, you know the voltage at the positive side of C6.? It is listed as 5.9VDC.? So you can infer that the voltage across C6 is 18.6V.? From there, you can find what the AC voltage across the winding should be, correct???
?
Also, in the schematic they show collector of Q9 at 6.61VDC and collector of Q4 at 5.9VDC.? But apparently Q4 and Q9 are tied together at the collector by wire 968.? What is the reason for the 0.6V drop?? Or is this an error???
?
-Frank
|
8513A T/R test set attenuator question
Greetings,
?
I have an 8510 VNA system running well with an 8515A S-parameter test set.
I'm trying to convert a spare 8511A (sampler only box) into an 8513A T/R test set.
That should just involve adding a 26 GHz directional bridge, a few hard lines + attenuators.
?
I found a few discrepancies in the 8513A Service Manual regarding installed attenuators.
Would a kind soul be able to confirm what is actually installed in the 8513A T/R test set?
A picture of the internals would be great start.
?
Thanks for the bandwidth,
Benoit Robert
?
?
?
|
Re: HP 141T Power Supply Help Needed
Make sure you also replace the gas voltage reference (used in low serial number instruments) with a zener.? My favorite is a TL431 plus a 60V-80V zener to stay within the TL431's max rating.? The zener has a large temperature coefficient, put it inside the
feedback loop.? This combo is more stable than a straight zener.? This can be a simple two-terminal board you tack across the tube without removing the tube.? See ?topic and photo album titled ¡°141T?reference?replacement".
A supply with gas reference outputs a high spike at turnon, which will trigger your crowbar.? If you look at the crowbar in the 8552B plugin, it has two thresholds and a timer, specifically to work around this.? It's ghastly, usually the HP engineers did better.
Dave Wise
toggle quoted message
Show quoted text
Dave,
Yes the sockets have caused intermittent connections.? The sockets are form Keystone Electronics P/N 4608 which are still available.? It appears that the brass contacts get annealed over time and the insertion force on one of the units I worked on had gone
to ZERO!? After taking the screws out of the collector, the transistor would have fallen out if it were not for the sill pad insulator.?? With new sockets the transistors fit tight!
Yes the voltage across C6 has to be measured at the cap because it is floating relative to ground.?
?I am thinking of designing a small board that will crowbar the -12.6 output if it goes beyond -14V.? I wonder if anyone would be interested in such a module?
Mike
-----Original Message-----
From: [email protected] < [email protected]> On Behalf Of Dave_G0WBX via groups.io
Sent: Friday, September 13, 2024 10:04 AM
To: [email protected]
Subject: Re: [HP-Agilent-Keysight-equipment] HP 141T Power Supply Help Needed
Take care with taking voltage measurements around the 141T PSU -12.6V regulator.
It is not "ground referenced" at the input.? It is essentially a floating Positive regulator with it's Positive output tied to "Ground", resulting in C6 "floating" at some arbitrary voltage on it's Positive side, though it's Negative side "Should" be at (or
very close to) the
-12.6 Output voltage.
To measure the raw DC voltage, you need to measure across C6, not from it's +ve terminal to Ground!
Re power transistor sockets.? Over the years I've always found them to be a major cause of all sorts of weirdness.? Best abandon them, as once the pin sockets have lost their spring temper, you'll never get a reliable contact between them and the power transistor
Base and Emitter leads.? Just fit the device conventionally, taking care with any needed insulation between the device case, and the heatsink, including insulating washers for the fixings etc.
If the collector of Q4 does get shorted to "Ground", the -12.6V will end up at an uncontrolled -ve voltage more(or less) equal to the raw voltage across C6.? Not good for some of the down stream loads, such as perhaps, the CRT flood gun filaments!? Especially
if there is no (or faulty) downstream crowbar protection, that should fire and "Blow" F4 if such a thing happens.
Apologies if I'm preaching to the converted, I'm not trying to teach anyone to suck eggs!? Just to raise awareness re the layout of that particular regulator circuit, as it is not entirely straightforward.
73.
Dave B.
--
Created on and sent from a Unix like PC running and using open source software:
|
Re: HP 141T Power Supply Help Needed
Dave,
Yes the sockets have caused intermittent connections. The sockets are form Keystone Electronics P/N 4608 which are still available. It appears that the brass contacts get annealed over time and the insertion force on one of the units I worked on had gone to ZERO! After taking the screws out of the collector, the transistor would have fallen out if it were not for the sill pad insulator. With new sockets the transistors fit tight!
Yes the voltage across C6 has to be measured at the cap because it is floating relative to ground.
I am thinking of designing a small board that will crowbar the -12.6 output if it goes beyond -14V. I wonder if anyone would be interested in such a module?
Mike
toggle quoted message
Show quoted text
-----Original Message----- From: [email protected] < [email protected]> On Behalf Of Dave_G0WBX via groups.io Sent: Friday, September 13, 2024 10:04 AM To: [email protected]Subject: Re: [HP-Agilent-Keysight-equipment] HP 141T Power Supply Help Needed Take care with taking voltage measurements around the 141T PSU -12.6V regulator. It is not "ground referenced" at the input. It is essentially a floating Positive regulator with it's Positive output tied to "Ground", resulting in C6 "floating" at some arbitrary voltage on it's Positive side, though it's Negative side "Should" be at (or very close to) the -12.6 Output voltage. To measure the raw DC voltage, you need to measure across C6, not from it's +ve terminal to Ground! Re power transistor sockets. Over the years I've always found them to be a major cause of all sorts of weirdness. Best abandon them, as once the pin sockets have lost their spring temper, you'll never get a reliable contact between them and the power transistor Base and Emitter leads. Just fit the device conventionally, taking care with any needed insulation between the device case, and the heatsink, including insulating washers for the fixings etc. If the collector of Q4 does get shorted to "Ground", the -12.6V will end up at an uncontrolled -ve voltage more(or less) equal to the raw voltage across C6. Not good for some of the down stream loads, such as perhaps, the CRT flood gun filaments! Especially if there is no (or faulty) downstream crowbar protection, that should fire and "Blow" F4 if such a thing happens. Apologies if I'm preaching to the converted, I'm not trying to teach anyone to suck eggs! Just to raise awareness re the layout of that particular regulator circuit, as it is not entirely straightforward. 73. Dave B. -- Created on and sent from a Unix like PC running and using open source software:
|
Re: HP 141T Power Supply Help Needed
Take care with taking voltage measurements around the 141T PSU -12.6V regulator.
It is not "ground referenced" at the input.? It is essentially a floating Positive regulator with it's Positive output tied to "Ground", resulting in C6 "floating" at some arbitrary voltage on it's Positive side, though it's Negative side "Should" be at (or very close to) the -12.6 Output voltage.
To measure the raw DC voltage, you need to measure across C6, not from it's +ve terminal to Ground!
Re power transistor sockets.? Over the years I've always found them to be a major cause of all sorts of weirdness.? Best abandon them, as once the pin sockets have lost their spring temper, you'll never get a reliable contact between them and the power transistor Base and Emitter leads.? Just fit the device conventionally, taking care with any needed insulation between the device case, and the heatsink, including insulating washers for the fixings etc.
If the collector of Q4 does get shorted to "Ground", the -12.6V will end up at an uncontrolled -ve voltage more(or less) equal to the raw voltage across C6.? Not good for some of the down stream loads, such as perhaps, the CRT flood gun filaments!? Especially if there is no (or faulty) downstream crowbar protection, that should fire and "Blow" F4 if such a thing happens.
Apologies if I'm preaching to the converted, I'm not trying to teach anyone to suck eggs!? Just to raise awareness re the layout of that particular regulator circuit, as it is not entirely straightforward.
73.
Dave B.
--
Created on and sent from a Unix like PC running and using open source software:
|
Hello Yves,
This second converter assembly is also used in the 8563E and also ESA
E4407B spectrum analyzers. In fact I think it is common for all
6.5/13.5/26.5GHz versions.
Mine had failed in the E4407B and it looks like it is related to an
incorrectly manufactured transformer inside the assembly.
I attached the Agilent Service Note related to this issue.
Part number affected is 5086-7958.
Thanks for the nice PDF with all explanations.
Regards,
Razvan
toggle quoted message
Show quoted text
On 13/09/2024 07:51, Yves Tardif via groups.io wrote: Hi Adri,
A friend of mine had a faulty 8595E, the problem comes from the option
card 130.
But I took the opportunity to look at the second converter, so here are
some details on this module.
Regards
Yves
|
Hi Adri, ? A friend of mine had a faulty 8595E, the problem comes from the option card 130. But I took the opportunity to look at the second converter, so here are some details on this module. ? Regards Yves ? ? hello Yves,
I received the 2nd converter and, after testing it on the bench (if you are interested I will give you the two test configurations) I installed it and voil¨¤, the machine is back to working like in the old days
It passes the amplitude and frequency calibrations and the confidence test without any problem
As it happened before the death of the YTO the analyzer still has a certain "deafness" after 5.5GHz
By virtue of the alignment of the spheres that I did a few weeks ago The defect is milder than it was then and I think it is due to a sub-optimal loss of insertion of the SYTF A3A2 As proof of the "deafness", while all the calibration routines, including the CAL YTF routine are performed without problems, the CAL MXR is not completed and gives the message CAL: HARMONICS INSUFFICIENT In fact I attach a screen of the 300MHz calibration signal from which it would seem that the very last harmonics are missing
Thank you in advance if can you send the analogous screen of your analyzer
In a few days I will receive a more reliable 6GHz RF generator than the current one and I will use it to try to perfect the alignment
Yves, I think that now you can reassemble and close your analyzer :-) Many thanks for all the measurements that you sent me and that were useful to trace the problem!
Adri
toggle quoted message
Show quoted text
----- Messaggio originale ----- Inviato: 08/09/2024 01:35:49 Oggetto: Re: [HP-Agilent-Keysight-equipment] HP 5086-7906 YTO
Hi Adri ? Here are the measurements connected to the first 2 converters as a reference. ? For the second converter I get a voltage of -1.678V instead of -3.5V as written in the service manual, I would say that we have similar voltages for this particular point. It seems far from the nominal point, but there may be an error in the manual. ? You are close to the goal, and this module may be repairable if there is no micro circuit, I have never opened a second converter in this series, but it is a possibility. ? When you open your second converter, send me pictures, I'm curious. ? Yves ? dear Yves, I didn't realize that you had made the measurements expressly to help me! I owe you then a double (even triple) thanks! Especially considering the fact that the front end of this instrument is really very dense, complicated and delicate! After my last email, in addition to doing the (essentially mechanical) intervention on the SYTF, I continued to check the input signal through the various sections by alternately injecting 300MHz @-20dBm and 5GHz @0 dBm The signals were found to be in specification up to the A3A9 bandpass (where the two paths join to form the intermediate frequency of 321.4MHz). At this point I could see the 5GHz signal (converted), present at that point at -47dBm (the specification says -43/-56dBm).
Instead, no trace of the 300MHz input signal
Investigatine?the second converter A3A4 in more depth
I checked?the signals at its input
On J1 there is 3.9214GHz at -46dBm? --->OK
On J3 there is 321.4MHz @ -37.7dBm? --->OK
On J4 I did not check 600MHz @-1dBm but the fact that the converter works in band 1 tells me that the oscillator is present and is in specification
The power supply is quite OK (5V are actually 4.75; 10V are 9.98 and the DC that controls the internal switch that should be +15V? (band 0) and -3.5V?(band 1) is actually +14.60V and -1.35V
This last value is rather far from the nominal and would have worried me BUT band 1 is exactly the one that works Conclusion: it would seem that the culprit is the second converter and in particular (probably) the internal switch that has the sole task of letting the signal? (which is already at 321.4MHz and does not need to be converted)?pass from its input to the output
In the next few days I will test the converter that meanwhile?has been?disassembled?; In the meantime I found a cheap one on the bay that should arrive in about ten days
Yves, I have a question: on the block diagram A3A4 is represented as a block that has 5 ports of which one (J5) would seem to be the connector that carries the power and the DC that controls the switch and that connects to A7 Analog Interface through the W35 cable
Actually,?this connector on the real component, does not have a name
Instead,?the name J5?, stamped on the case , indicates an SMB connector that has the writing LO TEST and in my analyzer?it is the one that?conveys the 600MHz LO signal? Now,?on the block diagram the connector that conveys the 600MHz signal has the name J4 (in my analyzer, the J4 connector was disconnected) This confusion of names and roles made me doubt that years ago, when I had disassembled this section I had not reassembled it correctly thanks again and best regards ----- Messaggio originale ----- Inviato: 06/09/2024 23:21:52 Oggetto: Re: [HP-Agilent-Keysight-equipment] HP 5086-7906 YTO
Hi Adri, ? I had sent you the measurements of the RF input-output mixer signals verification, I had disassembled the High Frequency section of my 8595E to be able to provide you with this information. If you have no more questions relating to this section, I will close my 8595E and do a flatness check to see if everything is correct. ? Yves ? ? thank you very much for all the suggestions and information with which I will continue my investigation
In the meantime I received the 2.5mm socket wrench and after?a little mod to adapt it?for the purpose (photo), I proceeded to align the spheres of the SYTF
I must say that the operation done under the microscope is quite simple
The nuts must be loosened (not too much) and with a small screwdriver that acts as a delicate lever, ?the supports of?the?spheres can be moved so that the coupling ring completely shadows them (i.e. the sphere is exactly in the center)
Removing the gold mesh was not necessary
The insertion loss has improved enormously:
at 5 GHz it is -6dBm at 6GHz it is -10dB (before the signal starting from this frequency was no longer visible) at 6.5 GHz it is only -3dB
I will install it when I have solved the problem of "blindness" in band 0 since the calibration constants that the instrument has in memory (in practice the tuning current data) are relative to the original SYTF and the "new" SYTW requires a? SYTF CAL is done
|
Re: Thermal Couple attached to heat sink
That's Thermocouple, to be precise.
Jim Ford, Laguna Hills, California, USA?
On Thu, Sep 12, 2024 at 1:20 PM, John Buie <johnrobertbuie@...> wrote:
On the A7A3 board in the 8663A? Sig Generator there is a heat sink with a slot that holds a thermal coupler. The slot is filled with thermal compound. Underneath the TC and the heat sink was a square insulator.? Mine has deteriorated beyond recognition.
Is the function of this just to press the TC against the heat sink, or is it also a thermal sink?? I¡¯m sure the same assembly is used on lots of HP boards.
Thanks,
John
|
Frank Mashockie