I'm putting together a toy box of various parts, as I plan to make a self contained Q-meter. I'm just?about to begin an RF sweep of the impedance converter / transformer?driver.
I used an isolated BNC chassis socket for the output, so I have the option of returning the zero-volt drive line to chassis inside the impedance converter / driver.
?I used two 2N3866s as I have several of them to hand.
The 50:1 transformer is a straight copy of the ON7CH one posted by him a few days ago (Thanks for the help, Guido!). It is built in a small offcut of 22mm copper water pipe and uses?50 turns on a T62-2 toroid..
?The single turn secondary is a piece of semi-rigid coax.
Pete G4GJL
On Sun, Sep 18, 2022 at 8:13 PM John KN5L < john@...> wrote: Hi Mikek,
The attenuators should not alter "Impedance Converter" output impedance.
The attenuators should have constant output impedance. Guessing
attenuator impedance is 50 ohm, based on R1 = 56.2 ohm. Steve, did you
measure input impedance of your Q1 Q2 demonstration board?
Today, we would use one of the several coax IC buffer amplifiers
available, for the "Impedance Converter Ass'y", as shown at bottom of:
"Q-Meter Prototype Schematic" using an AD8055.
John KN5L
On 9/18/22 12:53 PM, Mikek wrote:
> On Sat, Sep 17, 2022 at 03:06 PM, MAX wrote:
>
>? ? ?The output impedance of an emitter follower is the source impedance
>? ? ?driving the base divided by the current gain of the transistor plus
>? ? ?one.? This value is in parallel with any resistors that may be
>? ? ?connected to the emitter either DC or AC coupled.? So it goes back
>? ? ?to the impedance of the generator that is connected to J6.
>
> For those playing along, Here is a drop box schematic put together from
> the Oscillator to the injection transformer. I was thinking it may help
> to see the impedances.
> I don't have a clue what the attenuator does to the output impedance of
> the RF Power Amplifier.
> ?If I got something wrong notify me, and I'll correct it.
>
|
Hi Mikek,
The attenuators should not alter "Impedance Converter" output impedance.
The attenuators should have constant output impedance. Guessing
attenuator impedance is 50 ohm, based on R1 = 56.2 ohm. Steve, did you
measure input impedance of your Q1 Q2 demonstration board?
Today, we would use one of the several coax IC buffer amplifiers
available, for the "Impedance Converter Ass'y", as shown at bottom of:
"Q-Meter Prototype Schematic" using an AD8055.
John KN5L
toggle quoted message
Show quoted text
On 9/18/22 12:53 PM, Mikek wrote: On Sat, Sep 17, 2022 at 03:06 PM, MAX wrote:
The output impedance of an emitter follower is the source impedance
driving the base divided by the current gain of the transistor plus
one.? This value is in parallel with any resistors that may be
connected to the emitter either DC or AC coupled.? So it goes back
to the impedance of the generator that is connected to J6.
For those playing along, Here is a drop box schematic put together from
the Oscillator to the injection transformer. I was thinking it may help
to see the impedances.
I don't have a clue what the attenuator does to the output impedance of
the RF Power Amplifier.
?If I got something wrong notify me, and I'll correct it.
|
On Sat, Sep 17, 2022 at 03:06 PM, MAX wrote:
The output impedance of an emitter follower is the source impedance driving the base divided by the current gain of the transistor plus one.? This value is in parallel with any resistors that may be connected to the emitter either DC or AC coupled.? So it goes back to the impedance of the generator that is connected to J6.
For those playing along, Here is a drop box schematic put together from the Oscillator to the injection transformer. I was thinking it may help to see the impedances. I don't have a clue what the attenuator does to the output impedance of the RF Power Amplifier. ?If I got something wrong notify me, and I'll correct it. https://www.dropbox.com/s/gw07vsi32k3lkfx/HP4342A%20Injection%20transformer%20complete%20driver%20jpg%202.jpg?dl=0
|
Try calculating the output Z by measuring the chaange in outpuit V when terminated with small value resistor (of suitable power rating).
John KK6IL
toggle quoted message
Show quoted text
On 9/17/2022 9:49 AM, Steve Ratzlaff wrote: I don't know why my measurement gives the higher than expected output impedance. I've confirmed the (calibrated) VNWA accurately measures a 0.2 ohm resistor. My breadboard is "ugly style, soldered connections, over a copperclad PCB ground plane". The polarized caps are wired properly. The 2N5109's are modern Central Semi.
Steve
|
I am a little surprised that HP didn't put a .01 uF ceramic capacitor in parallel with that 22 uF electrolytic. Maybe it is a physically small tantalum?
73 (Regards).
Max K 4 O D S.
I've Never Lost the Wonder.
toggle quoted message
Show quoted text
-----Original Message----- From: [email protected] [mailto: [email protected]] On Behalf Of Tom Lee Sent: Saturday, September 17, 2022 12:17 AM To: [email protected]Subject: Re: [Test Equipment Design & Construction] Making a Q-meter / Your analysis is spot on (well done for an FPGA designer!); it's the same as that in my first post, although I used 100 for beta of both transistors. In any case, the intrinsic output impedance is very low. The actual output impedance will be dominated by parasitic resistances. For the 2N5109, I would guess emitter resistances of a few hundred milliohms. That would be in series with whatever ESR the output cap possesses. Overall, I would expect something under an ohm. Measurements much greater than that would lead me to check the capacitor, wiring parasitics, and VNA cal. --Tom -- Prof. Thomas H. Lee Allen Ctr., Rm. 205 350 Jane Stanford Way Stanford University Stanford, CA 94305-4070 On 9/16/2022 22:04, John Kolb wrote:
Googling Output resistance of an emitter follower, I find it's input
resistance divided by (Beta + 1) plus some other factors, internal
emitter resistance of Q2, etc, that I as a FPGA designer couldn't
understand. Beta of a darling configuration is the beta of Q1 times
beta of Q2. 50 ohm Rin (R1, R2, R3 in parallel) /(100 (beta of Q1) *
30 (beta of Q2)) would be 0.018 ohms. Beta numbers are SWAG. Would a
real analog engineer speak up here? At any rate, the output
resistance would seem to be much lower than R5, 220 ohms, or 75 ohms,
R1 of transformer.
Would really like to read the complete manual for the 4342A.
|
I may be premature in answering this question because it may have already been answered. ? The output impedance of an emitter follower is the source impedance driving the base divided by the current gain of the transistor plus one.? This value is in parallel with any resistors that may be connected to the emitter either DC or AC coupled.? So it goes back to the impedance of the generator that is connected to J6.? Then the output impedance is multiplied by the turns ratio squared of the transformer.? ? 73 (Regards). ? Max K 4 O D S. ? I've Never Lost the Wonder. ?
toggle quoted message
Show quoted text
From: [email protected] [mailto:[email protected]] On Behalf Of Mikek
Sent: Friday, September 16, 2022 7:51 PM
To: [email protected]
Subject: Re: [Test Equipment Design & Construction] Making a Q-meter / ? I'm not sure I agree, the schematic I posted shows a 75¦¸ resistor across the transformer as Zo.
And I'm waiting for an analysis of the impedance converter amp to see if we can find it's output impedance.
I'm speculating that it is low, but the 220¦¸ resistor has me questioning.
????????????????????? Mikek
|
Lots of 4342A manuals floating around - don't know why I had trouble finding one before.
<> Mar 1983 version 04342-90009 contains only through section 5, thus no schematics.
XDEVS also has version 04342-99004 Nov 1975 which has MUCH better photos. <>
That one is also available at <>
A complete 04342-90009 Mar 1983 can be found at <> Thanks Steve/
John
toggle quoted message
Show quoted text
On 9/17/2022 6:26 AM, Jeff Green wrote: This is the manual for the HP 4342A:
I hope it is useful.
_._,_._,_
------------------------------------------------------------------------
|
Much better! I believe those numbers a lot more than the previous ones.
And a low-MHz corner frequency (where the inductive component of the output impedance equals the resistive one) is about where it's expected, too.
A tip of the hat to you for your persistence, Steve.
-- Cheers,
Tom
--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
toggle quoted message
Show quoted text
On 9/17/2022 11:50, Steve Ratzlaff wrote: I finally have much better results! The Q1 input transistor makes all the difference. I've been trying various Q1's and have found several that work very well. Low end input impedance is now under one ohm. The same 2N5109 Q2 is used; I haven't tried changing it (Central Semi, DC beta 110). I've been looking at the 1 MHz point for my Q1 comparisons. Best is 1.05 ohms with an obsolete Motorola 2N4401; a metal can BC108B gives 1.07 ohms; a metal can 2N3947 gives 1.10 ohms; an obsolete Motorola PN2222A gives 1.11 ohms.
50 kHz 0.86 ohms input impedance
1 MHz 1.05 ohms
10 MHz 2.93 ohms
20 MHz 5.33 ohms
30 MHz 7.79 ohms
Steve
On 9/17/2022 2:58 AM, Tom Lee wrote:
If one performs a more detailed analysis, a good approximation for the output resistance should be:
(rb2/beta2) + re2 + (1/gm2) + ESR
a grand total of an ohm or so. Closer to your measured value, but...
--Tom
|
That should be "output impedance".??? :)
toggle quoted message
Show quoted text
On 9/17/2022 11:50 AM, Steve Ratzlaff via groups.io wrote: I finally have much better results! The Q1 input transistor makes all the difference. I've been trying various Q1's and have found several that work very well. Low end input impedance is now under one ohm. The same 2N5109 Q2 is used; I haven't tried changing it (Central Semi, DC beta 110). I've been looking at the 1 MHz point for my Q1 comparisons. Best is 1.05 ohms with an obsolete Motorola 2N4401; a metal can BC108B gives 1.07 ohms; a metal can 2N3947 gives 1.10 ohms; an obsolete Motorola PN2222A gives 1.11 ohms.
50 kHz 0.86 ohms <output> impedance
1 MHz 1.05 ohms
10 MHz 2.93 ohms
20 MHz 5.33 ohms
30 MHz 7.79 ohms
Steve
On 9/17/2022 2:58 AM, Tom Lee wrote:
If one performs a more detailed analysis, a good approximation for the output resistance should be:
(rb2/beta2) + re2 + (1/gm2) + ESR
I don't know the parasitic resistances of the transistor, unfortunately, so I'm forced to guess. But a good RF transistor should not have large base resistances. I'll make up a number, 25 ohms, which is higher than I think is reasonable, partly to compensate for neglecting Q1's contribution to the base resistance of Q2 (a few ohms). For beta2, I'll use the universal guess of 100, so the first term contributes 0.25 ohms. You can substitute your own values in the equation above.
The parasitic emitter resistance should similarly be small; I'll conservatively guess 0.25 ohms for the 2N5109. The actual value is likely half that.
At a current of 84mA, 1/gm2 is around 0.3 ohms.
You measured a capacitor ESR of 0.26 ohms, so adding up all the terms gives us a grand total of an ohm or so. Closer to your measured value, but...
--Tom
|
I finally have much better results! The Q1 input transistor makes all the difference. I've been trying various Q1's and have found several that work very well. Low end input impedance is now under one ohm. The same 2N5109 Q2 is used; I haven't tried changing it (Central Semi, DC beta 110). I've been looking at the 1 MHz point for my Q1 comparisons. Best is 1.05 ohms with an obsolete Motorola 2N4401; a metal can BC108B gives 1.07 ohms; a metal can 2N3947 gives 1.10 ohms; an obsolete Motorola PN2222A gives 1.11 ohms.
50 kHz 0.86 ohms input impedance
1 MHz 1.05 ohms
10 MHz 2.93 ohms
20 MHz 5.33 ohms
30 MHz 7.79 ohms
Steve
toggle quoted message
Show quoted text
On 9/17/2022 2:58 AM, Tom Lee wrote: If one performs a more detailed analysis, a good approximation for the output resistance should be:
(rb2/beta2) + re2 + (1/gm2) + ESR
I don't know the parasitic resistances of the transistor, unfortunately, so I'm forced to guess. But a good RF transistor should not have large base resistances. I'll make up a number, 25 ohms, which is higher than I think is reasonable, partly to compensate for neglecting Q1's contribution to the base resistance of Q2 (a few ohms). For beta2, I'll use the universal guess of 100, so the first term contributes 0.25 ohms. You can substitute your own values in the equation above.
The parasitic emitter resistance should similarly be small; I'll conservatively guess 0.25 ohms for the 2N5109. The actual value is likely half that.
At a current of 84mA, 1/gm2 is around 0.3 ohms.
You measured a capacitor ESR of 0.26 ohms, so adding up all the terms gives us a grand total of an ohm or so. Closer to your measured value, but...
--Tom
|
Appended to bottom of:
is a Q-Meter prototype schematic proposal.
For greatest accuracy, RF to injection transformer should have low
impedance, proposing an AD8055 direct connection.
A requirement is a AWG with calibrated output for RF In.
Assumption: "RF Voltmeter Prototype" will behave appropriately when
placed onto a proper PWB. Suggestion J111 or J112 input FET.
Can envision two PWBs. One large with all components. Second smaller,
mezzanine board, above the 50:1 injection transformer, for L and C
connection points.
John KN5L
|
On Sat, Sep 17, 2022 at 09:49 AM, Steve Ratzlaff wrote:
I don't know why my measurement gives the higher than expected output
impedance. I've confirmed the (calibrated) VNWA accurately measures a
0.2 ohm resistor. My breadboard is "ugly style, soldered connections,
over a copperclad PCB ground plane". The polarized caps are wired
properly. The 2N5109's are modern Central Semi.
Steve
---- Do you have a local bypass cap across the supplies? Preferably one large, one small cap in parallel. Ozan
|
If you don't want to install yet another program, will do it in a web browser.? At least it did for me in Firefox on a Mac.
Very slow if you have it load the whole document in "continuous scroll view mode" - leave it in "single page view mode".
I believe the page of interest is 109 - type it into the page range box at the bottom.
Orin.
On Sat, Sep 17, 2022 at 7:32 AM Mikek < amdx@...> wrote: On Sat, Sep 17, 2022 at 06:26 AM, Jeff Green wrote:
Hey Jeff, There is a more complete service manual on the Bama site. I had to download a djvu reader to open it, but once I did that,
it's pretty nice!
This reader is what I downloaded.
??????????????? Mikek_._,_._,
|
I don't know why my measurement gives the higher than expected output impedance. I've confirmed the (calibrated) VNWA accurately measures a 0.2 ohm resistor. My breadboard is "ugly style, soldered connections, over a copperclad PCB ground plane". The polarized caps are wired properly. The 2N5109's are modern Central Semi.
Steve
toggle quoted message
Show quoted text
On 9/17/2022 2:58 AM, Tom Lee wrote: If one performs a more detailed analysis, a good approximation for the output resistance should be:
(rb2/beta2) + re2 + (1/gm2) + ESR
I don't know the parasitic resistances of the transistor, unfortunately, so I'm forced to guess. But a good RF transistor should not have large base resistances. I'll make up a number, 25 ohms, which is higher than I think is reasonable, partly to compensate for neglecting Q1's contribution to the base resistance of Q2 (a few ohms). For beta2, I'll use the universal guess of 100, so the first term contributes 0.25 ohms. You can substitute your own values in the equation above.
The parasitic emitter resistance should similarly be small; I'll conservatively guess 0.25 ohms for the 2N5109. The actual value is likely half that.
At a current of 84mA, 1/gm2 is around 0.3 ohms.
You measured a capacitor ESR of 0.26 ohms, so adding up all the terms gives us a grand total of an ohm or so. Closer to your measured value, but...
--Tom
|
Is it possible to come up with a formula or spreadsheet that would calculate Q from 3db,
with the caveat that instead of using 0.707, use the number 0.7 and normalize it to 3db?
That would be great for needle type meters. For oscilloscopes, I'd like to see
0.714 normalized to 3db, why because I like to peak at 7 units and raise/lower the frequency to 5 units.
5 / 7 = 0.714.?
????????????????????????? Thanks, Mikek
?????
|
On Sat, 17 Sep 2022 at 17:02, John KN5L < john@...> wrote: Hi Mikek,
Revisiting with my V-Meter prototype as shown in:
DUT = T50-2 37T #24 with 330pF C0G
Different L C components. VNWA BW Q measurement = 211 for these parts.
Peaked to VOM full scale = 1V out of fixture, AWG = 204mV
Q = 50 / .204 = 245? (3.3387 MHz peak)
3dB down = 0.7079
Best I can resolve with VOM is 0.71, measuring two frequencies;
3.3317 and 3.3456 MHz with difference = 0.0139 MHz
Q = 3.3387 / 0.0139 = 240
Nice Match!
Which then questions, which is correct VNWA BW or prototype Voltmeter
with injection transformer?
John
Almost certainly neither is correct.??? The challenge will be to determine the uncertainty of each method, but even for simple instruments, working out the uncertainty is a non-trivial process.?
It would be interesting to put your VNA bandwidth measurements through the code in the big (100+ page) document from NPL about Q measurement with a VNA. That doesn¡¯t just use 3 points, but multiple points.?
The NPL document is in the files section of the group.?
Dave? -- Dr. David Kirkby, Kirkby Microwave Ltd, drkirkby@...Telephone 01621-680100./ +44 1621 680100 Registered in England & Wales, company number 08914892. Registered office: Stokes Hall Lodge, Burnham Rd, Althorne, Chelmsford, Essex, CM3 6DT, United Kingdom
|
On Sat, Sep 17, 2022 at 09:02 AM, John KN5L wrote:
Which then questions, which is correct VNWA BW or prototype Voltmeter
with injection transformer?
That's what my 3db request was getting at. ?????????????????? Thank, Mikek
|
Hi Mikek,
Revisiting with my V-Meter prototype as shown in:
DUT = T50-2 37T #24 with 330pF C0G
Different L C components. VNWA BW Q measurement = 211 for these parts.
Peaked to VOM full scale = 1V out of fixture, AWG = 204mV
Q = 50 / .204 = 245 (3.3387 MHz peak)
3dB down = 0.7079
Best I can resolve with VOM is 0.71, measuring two frequencies;
3.3317 and 3.3456 MHz with difference = 0.0139 MHz
Q = 3.3387 / 0.0139 = 240
Nice Match!
Which then questions, which is correct VNWA BW or prototype Voltmeter
with injection transformer?
John
toggle quoted message
Show quoted text
On 9/17/22 9:46 AM, Mikek wrote: Can you do a 3db Q measurement on the DUT. Set resonance a 1 unit then
raise and lower the frequency until the meter reads 0.707 and record the
frequencies.
High frequency minus low frequency = xxx then resonant frequency divided
by xxx = Q
Just wondering what the real Q is.
????????????????????? Thanks, Mikek
|
Mikek, The HP-400E measured DUT uses same components as in the Prototype
meter section. Within that section is "VNWA S21 Q Measurement" with Q = 229.
John KN5L
toggle quoted message
Show quoted text
On 9/17/22 9:46 AM, Mikek wrote: Can you do a 3db Q measurement on the DUT. Set resonance a 1 unit then
raise and lower the frequency until the meter reads 0.707 and record the
frequencies.
High frequency minus low frequency = xxx then resonant frequency divided
by xxx = Q
Just wondering what the real Q is.
????????????????????? Thanks, Mikek
|
Mikek, The SimSmith model is assuming an Ideal transformer, with no
loss. The 50 points is a function of Generator 50 Ohm output impedance.
Lower generator output impedance will reduced Q loss. A possible
solution is an AD8055 amplifier at the injection transformer.
I'm not sure how to measure high frequency rolloff, as it's a function
of coefficient of coupling. Rather hard to measure a 2500:1 Z ratio
coefficient of coupling.
Low frequency rolloff is a function of inductance, T50-43 50T = 1100uH,
will have a rather low rolloff frequency.
John KN5L
toggle quoted message
Show quoted text
On 9/17/22 9:34 AM, Mikek wrote: John, what are the low and high frequency rolloff points of your 50 to 1
#43 material transformer?
?Is some of that missing 50 points of Q, caused by the loss in the
transformer primary?
?What is the Q of your 50¦¸ primary
What do you think are the other losses?
?
???????????????????? Thanks, Mikek
|
Pete_G4GJL
John Kolb
Dr. David Kirkby, Kirkby Microwave Ltd