Re: HP 141T Power Supply Help Needed
Hi guys,
Good points here. What happens when you place a cap with zero ESR
at the output of a linear power supply is that you intrude on the phase margin
and the loop becomes unstable. This can easily be verified by running an
LTspice simulation.
?
The remedy would be to add a small series resistor to the cap. I have seen this
done in the hp8640B signal generator power supply - a long time ago.
?
What is worth noting is that the output cap is part of the load and not part of the loop.
?
G?ran
|
Re: Keysight customer support policy
<clip>
Our support team forwarded your inquiry to the contact center in Finland.
Kindly note that Keysight products are designed, manufactured and tested for professional and industrial use. They are not designed or tested for use by Consumers and therefore we are not able to supply them or provide service for them in case you are a private individual.
In case you work for a company, kindly provide us the company name and address as well as your company email address and we can continue working with your inquiry.
<clip>
?
Anyone with ideas about one (the sensor) or the other (bypass the policy-barrier)?
I'm happy to give the details Keysight ask for once I get in contact with a real and friendly person.
Obviously the sensor is more interesting now, but I have one other Agilent component to ask about.
You work as a freelance consultant. It might help to have a web page you can point them to.
?
|
Keysight customer support policy
Hi group,
?
I tried asking Keysight support for possibility to repair or buy spare parts for a HP E4413A 26.5 GHz power sensor and giving the serial number and my full name.
?
The sensor works fine, but the 3.5mm connector is suspect.
Both the center pin and the ground connection can easily be rotated.
Not sure this is a standard feature or potential problem so I asked support.
?
However, I seem to end up at a AI-generated door guard asking me to register whatever reply I give.
I don't find it very meaningful to start registering before I even know if I have a problem or not.
At first I think I got in contact with the US support, but after few unfruitful trials they sent me to the Finnish representative repeating the :
?
<clip>
Our support team forwarded your inquiry to the contact center in Finland.
Kindly note that Keysight products are designed, manufactured and tested for professional and industrial use. They are not designed or tested for use by Consumers and therefore we are not able to supply them or provide service for them in case you are a private individual.
In case you work for a company, kindly provide us the company name and address as well as your company email address and we can continue working with your inquiry.
<clip>
?
Yes, I'm a retired professional with long experience dealing with/buying from HP and Agilent, and with my own quite inactive design company, but how to ask a fair question without filling my "shoe-number" first?
I value privacy and respect.
?
Anyone with ideas about one (the sensor) or the other (bypass the policy-barrier)?
I'm happy to give the details Keysight ask for once I get in contact with a real and friendly person.
Obviously the sensor is more interesting now, but I have one other Agilent component to ask about.
?
Thanks!
?
|
Re: Crystal detector with 10dB attenuator in series - is it for matching purposes?
Excellent answers, thanks all.
|
Re: HP 141T Power Supply Help Needed
To amplify your comment, many times older linear supplies, when they are brute force recapped, with the best, long life, low esr caps you can find, will become beacons of chaotic oscillation. I put a low ESR ceramic capacitor in the feedback loop of a unity gain OPAMP circuit the other day, and was amazed that it went totally unstable. It was an older LM301 variety, where the response characteristics of the opamp are a blank slate until the designer adds in the stability network. The designer's efforts were wiped out by my ceramic capacitor replacement for his 1uf back-to-back emulation of a non-polar electrolytic. Sometimes the modern best replacement is not suitable for use in an older instrument. -Chuck Harris On Mon, 2 Sep 2024 22:40:20 +0100 "Dave_G0WBX via groups.io" <g8kbvdave@...> wrote: Hi.
It's just a generic term (damping is another) to slow down the
response of a loop, to prevent it going unstable.
Not an uncommon issue in some linear PSU's like those -12.6V or -100V
regulator topologies that are not "conventionally" laid out.
Unless you 'scope the thing, you often don't know it's gone unstable,
other than some passive parts sometimes cook for no obvious reason.?
(That R49 for example!? See Mike's comment in another mail.)
It is also not uncommon, when modern transistors are used to replace
older failed parts (that are not available any more) and the new
device has "a lot" more gain at HF than the originals, that regulator
or other control loops suddenly become "unruly" under some operating
conditions.
Sometimes a low value "Base Stopper" resistor (a few Ohms, or a
ferrite bead) in series with the new device's Base connection can
restore sanity!
Take care.
Dave B.
|
Re: HP 141T Power Supply Help Needed
Thanks, Dave B.? "To slug" = "to make sluggish".? Thought so.
Still, I still have never heard the term in 50-plus years of electronics work, including college courses in control system theory.
The base stopper idea is great, but when I have to replace e.g. 2N3055 in a power supply, I also look for a place I can add a Miller capacitor.
Dave W
toggle quoted message
Show quoted text
Hi.
It's just a generic term (damping is another) to slow down the response
of a loop, to prevent it going unstable.
Not an uncommon issue in some linear PSU's like those -12.6V or -100V
regulator topologies that are not "conventionally" laid out.
Unless you 'scope the thing, you often don't know it's gone unstable,
other than some passive parts sometimes cook for no obvious reason.?
(That R49 for example!? See Mike's comment in another mail.)
It is also not uncommon, when modern transistors are used to replace
older failed parts (that are not available any more) and the new device
has "a lot" more gain at HF than the originals, that regulator or other
control loops suddenly become "unruly" under some operating conditions.
Sometimes a low value "Base Stopper" resistor (a few Ohms, or a ferrite
bead) in series with the new device's Base connection can restore sanity!
Take care.
Dave B.
--
Created on and sent from a Unix like PC running and using open source software:
|
Re: Schottky Diode degradation / Replacement in instrument Sram Battery backup circuits ?
Is this failure is related to having a few uA running through the metal junction for decades?
Personally, if it took 30+ years to get so leaky, then I would still replace it with a new schottky - and a new battery.?
?
?
|
Re: Instrument Sram ( non Dallas ) Battery replacement: Original 3.0V Li+ >> To 3.6V Li+ & Schottky Diodes ?
?
The question is about substituting? a long life, low leakage 3.6V Li+ Solder Tab, 10+ year battery
?
???? ( Such as Tadiran or SAFT LSH-14 cell ) .... of at least 1200 mAh to 1600mAh capacity,
?
instead of
??? the original 3.0V Li+ Solder Tab BR-2/3a? long life, low leakage cell,
?
for the battery backup of Cal Data in Sram, in a lab instrument.
?
?
?
?
?
?
|
Re: Crystal detector with 10dB attenuator in series - is it for matching purposes?
Hi all Yeas its a very nice tool.Mine came with
High Frequency Swept Measurements. Hardy ? ? Very good! This what I was getting at.? I have a little slide rule from HP that does the 20 LOG(1+rho1*rho2)^2? for the mismatch. Nice calculation tool. Helps when relating VSWR effects to amplitude ripple, especially in big system, where I started 50 years ago. I realize generator return loss degrades at higher frequency (what doesnt?).? J. Kruth
In a message dated 9/2/2024 6:44:52 PM Eastern Standard Time, jmrhzu@... writes:
If it helps to add some numbers to all this then consider the case up around 22GHz where the sig gen source VSWR might be 1.8:1 typical.? This is poor enough to add significant mismatch uncertainty even if you connected it to a decent 26GHz power meter to calibrate the level. The sensor VSWR might be 1.2:1 at this frequency. The poor source VSWR of the signal generator would give an overall mismatch uncertainty of +/- 0.23dB for the signal level calibration before it gets connected to the HP 8473C. If the input VSWR of the 8473C is 2:1 at 20GHz then the mismatch uncertainty is going to be about +/- 0.8dB when the sig gen is connected to it. However, if a decent isolating attenuator is used for the calibration stage, the source VSWR of the sig gen might be improved to 1.2:1.? Therefore, the calibration phase using the power meter would benefit from having the mismatch uncertainty reduced to about +/- 0.07dB. When the connection is then made to the 8473C the mismatch uncertainty (with the isolating attenuator ahead of it) should reduce to about +/- 0.27dB. In both cases, adding the isolation attenuator provides a substantial improvement in mismatch uncertainty.
|
Re: Crystal detector with 10dB attenuator in series - is it for matching purposes?
Very good! This what I was getting at.?
I have a little slide rule from HP that does the 20 LOG(1+rho1*rho2)^2? for the mismatch. Nice calculation tool. Helps when relating VSWR effects to amplitude ripple, especially in big system, where I started 50 years ago. I realize generator return loss degrades at higher frequency (what doesnt?).? J. Kruth
In a message dated 9/2/2024 6:44:52 PM Eastern Standard Time, jmrhzu@... writes:
?
If it helps to add some numbers to all this then consider the case up around 22GHz where the sig gen source VSWR might be 1.8:1 typical.?
?
This is poor enough to add significant mismatch uncertainty even if you connected it to a decent 26GHz power meter to calibrate the level. The sensor VSWR might be 1.2:1 at this frequency.
The poor source VSWR of the signal generator would give an overall mismatch uncertainty of +/- 0.23dB for the signal level calibration before it gets connected to the HP 8473C.
?
If the input VSWR of the 8473C is 2:1 at 20GHz then the mismatch uncertainty is going to be about +/- 0.8dB when the sig gen is connected to it.
?
?
However, if a decent isolating attenuator is used for the calibration stage, the source VSWR of the sig gen might be improved to 1.2:1.? Therefore, the calibration phase using the power meter would benefit from having the mismatch uncertainty reduced to about +/- 0.07dB.
?
When the connection is then made to the 8473C the mismatch uncertainty (with the isolating attenuator ahead of it) should reduce to about +/- 0.27dB.
?
In both cases, adding the isolation attenuator provides a substantial improvement in mismatch uncertainty.
?
?
|
Re: Crystal detector with 10dB attenuator in series - is it for matching purposes?
If it helps to add some numbers to all this then consider the case up around 22GHz where the sig gen source VSWR might be 1.8:1 typical.?
?
This is poor enough to add significant mismatch uncertainty even if you connected it to a decent 26GHz power meter to calibrate the level. The sensor VSWR might be 1.2:1 at this frequency.
The poor source VSWR of the signal generator would give an overall mismatch uncertainty of +/- 0.23dB for the signal level calibration before it gets connected to the HP 8473C.
?
If the input VSWR of the 8473C is 2:1 at 20GHz then the mismatch uncertainty is going to be about +/- 0.8dB when the sig gen is connected to it.
?
?
However, if a decent isolating attenuator is used for the calibration stage, the source VSWR of the sig gen might be improved to 1.2:1.? Therefore, the calibration phase using the power meter would benefit from having the mismatch uncertainty reduced to about +/- 0.07dB.
?
When the connection is then made to the 8473C the mismatch uncertainty (with the isolating attenuator ahead of it) should reduce to about +/- 0.27dB.
?
In both cases, adding the isolation attenuator provides a substantial improvement in mismatch uncertainty.
?
?
|
Re: HP 141T Power Supply Help Needed
Hi.
It's just a generic term (damping is another) to slow down the response of a loop, to prevent it going unstable.
Not an uncommon issue in some linear PSU's like those -12.6V or -100V regulator topologies that are not "conventionally" laid out.
Unless you 'scope the thing, you often don't know it's gone unstable, other than some passive parts sometimes cook for no obvious reason.? (That R49 for example!? See Mike's comment in another mail.)
It is also not uncommon, when modern transistors are used to replace older failed parts (that are not available any more) and the new device has "a lot" more gain at HF than the originals, that regulator or other control loops suddenly become "unruly" under some operating conditions.
Sometimes a low value "Base Stopper" resistor (a few Ohms, or a ferrite bead) in series with the new device's Base connection can restore sanity!
Take care.
Dave B.
--
Created on and sent from a Unix like PC running and using open source software:
|
Re: 8566B Can't Calibrate
I fully agree. Subbing a 10dB attn sounds like a plan to me. Unfortunately I don't have time this week, but perhaps next I'll give it a go. At any rate, it's clearly the first thing I should try.
|
Jack and Bruce - many thanks. It all helps to work out what's happening.
I only posted a photo of what I saw with the default calibration where the SA is automatically set at 10dB. It's the exact same image I'm seeing when I force it into 0dB setting. So I'm getting nothing with 0 and 10dB. I've now taken pictures of what I see on the other settings from 20 to 70dB. I did try to describe these in another post, but a pictures worth a thousand words so I've attached pix for 20, 30, 40, 50, 60 and 70dB attenuation. Enjoy...
|
Re: Crystal detector with 10dB attenuator in series - is it for matching purposes?
On Mon, Sep 2, 2024 at 09:53 PM, Jeff Kruth wrote:
Some of the responses did not answer the question you asked. To answer the question below: modern signal generators, like the 8340, have a leveling loop to maintain a preset level. If you investigate to practical result of using this, it basically gives an infinitely low generator VSWR (very high return loss, good match) as no external load can cause significant amplitude ripple due to mismatch, as the loop automatically corrects this. I learned this from an HP APP note many years ago..
In my experience, and this is across many decades of using all kinds of lab sig gens, the signal generators that use a classic levelling loop do not give an infinitely low generator VSWR up at high RF frequencies. This is because a compromise has to be made between the source defining resistor at the output of the levelling loop and the added ESR of all the attenuator switches and other connections/circuitry after the levelling loop. The source VSWR is often quite good at low frequencies but the ESR (and general mismatch) of everything after the levelling loop creeps up with increasing frequency and this can spoil the source match at the output port of the sig gen.
?
Usually, the source impedance will be have a VSWR spec of about 1.5:1 at higher generator output levels because of this. For sig gens that operate up to about 20GHz, the source VSWR may climb as high as 2:1 in some cases.
?
Once the attenuator sections get used for lower output levels, the source match should improve in most cases.
|
Re: Crystal detector with 10dB attenuator in series - is it for matching purposes?
Some of the responses did not answer the question you asked. To answer the question below: modern signal generators, like the 8340, have a leveling loop to maintain a preset level. If you investigate to practical result of using this, it basically gives an infinitely low generator VSWR (very high return loss, good match) as no external load can cause significant amplitude ripple due to mismatch, as the loop automatically corrects this. I learned this from an HP APP note many years ago..
However, lets look at the return loss of a 10 dB pad with a short circuit on the end. It should have a 20 dB return loss due to the short, but in reality, it will be less due to imperfections in the actual attenuator performance. Still, adding a pad in gives a better VSWR to the crystal detector, regardless of the detector performance.
So, as to you original question, could I use a different (lower) value pad, yes, you could. It may alter this peak to peak observed ripple over a broadband but this will probably be small anyway if quality devices are used.
One caveat is that the best diode detector performance is at a level of about -20 dBm or so. If you want the best logarithmic performance, a pad may be used to reduce incident levels to achieve this, as well as provide improved broadband VSWR performance. Hope this helps. Jeff Kruth
In a message dated 9/2/2024 2:19:18 PM Eastern Standard Time, saevartj@... writes:
?
All, thanks for the replies.
?
Does an RF signal gen such as say 8340B generally have poor source impedance? Has someone examined this before?
?
I am wondering if the problem of mismatch and standing waves is a result of a poor sig gen impedance as jmr says or a poor input impedance of the diode detector.
|
Re: Crystal detector with 10dB attenuator in series - is it for matching purposes?
If it helps, this is the setup I typically use to level a microwave signal source and provide a low source VSWR. This provides levelling to the accuracy of the power meter and the source impedance is defined by the relevant series 50 ohm resistor in the twin resistor splitter.?
?
In this case the DUT (device under test) would be the HP 8473C diode. The levelling system below should provide an ultra low source VSWR and quite accurate levelling. Correction factors vs frequency can be added for the power sensor efficiency if required.
?
I usually automate all this via GPIB and it works really well up to the frequency limits of my test gear here.?
?
Because external levelling is used, it doesn't really matter (within reason) how lossy the coax cable is at the output of the synthesised sweeper.
?
?
?
|
Re: 8566B Can't Calibrate
Never mind my previous question - the picture you sent confirms the problem is most likely the attenuator.? I like the idea of subing a 10dB fixed for the switched attenuator followed by RECALL 8 and then 9.? This would confirm an attenuator problem.? If you go directly to replacing the switched attenuator, what is to say that it too may have a problem (unless you confirm B4 installation)
?
Cheers!
?
Bruce
|
One more question and then I'll "buzz off" and watch.? Did you see anything when you did RECALL 8.
?
I read that you did not see anything when you did RECALL 9
?
If RECALL 8 worked and RECALL 9 did not,? ?The problem is possibly the adjustment of the 18.4 MHz oscillator on A4A5.??
?
If neither RECALL 8 or 9 work, the attenuator is likely.? The suggestion of substituting a 10dB fixed attenuator for the variable attenuator will confirm.? Once you have substituted the 10dB, try RECALL 8 and 9.
?
If they work, it is VERY likely the attenuator.? There are videos on line and significand tata on this site about how to overhaul the attenuator.? It is a serious pain, but the results are worth the effort.
?
Cheer!
?
Bruce
?
|
Re: Crystal detector with 10dB attenuator in series - is it for matching purposes?
These have a DC output, not RF so input impedance of the test equipment doesn't matter. They typically feed the vertical a scope or DC? voltmeter .
The input VSWR of the 8473C is quite good up to about 18GHz, but above this a decent attenuator will have a lower VSWR.? So the attenuator acts to isolate the less than perfect VSWR of the sig gen from the less than perfect VSWR of the 8473C detector.
However, a lot depends on what frequency range you want to operate up to.
?
The source impedance of a typical sig gen will improve at lower power levels when using its internal attenuator so the sig gen might have a fairly good source match when delivering -10dBm. I've not measured the 8340B so I can't comment on this.
?
However, it is unlikely to be as good as a decent 3.5mm 10dB attenuator that is rated to 26.5GHz. Every little bit helps when trying to minimise mismatch uncertainty...
?
If you ultimately want to use the 8473C for external levelling of the sig gen then maybe consider using a 11667B splitter to provide the feedback path as this will have very low source VSWR when the external levelling is in closed loop.
|
G?ran Krusell
Dr. David Kirkby, Kirkby Microwave Ltd
Jinxie
Michael A. Terrell