Hi David,

Thanks, I had seen that design on the ARF forum and was originally considering going that route but then had trouble finding the BFG35 transistors.? I was thinking that since the newer higher output MMIC's were readily available it might be a better/simpler design to go with the MMIC and reduce the parts count quite a bit.? Hopefully the RF choke and biasing wont complicate this too much for the lower frequencies.

Thanks
DB


"This design may help with avoiding re-inventing the wheel:

?

?

(No personal experience, just remember the conversation when it occurred back on the AntiqueRadioForums)

?

David

?

1GHz only?

On Mar 21, 2022, at 19:52, n8zmTWH via groups.io <n8zmTWH@...> wrote:

?

What Frequencies?

?

?

Tom, N8ZM

?

From: [email protected] <[email protected]> On Behalf Of Lothar baier
Sent: Monday, March 21, 2022 8:01 PM
To: [email protected]
Subject: Re: [HP-Agilent-Keysight-equipment] Protecting Equipment

?

Agilent E8257D 20GHz Signal generator ?, N9030A 50GHz spectrum analyzer , power sweep in 0.5Db steps from 0 to 30dBm CW

?

From: [email protected] <[email protected]> On Behalf Of n8zmTWH via groups.io
Sent: Monday, March 21, 2022 6:48 PM
To: [email protected]
Subject: Re: [HP-Agilent-Keysight-equipment] Protecting Equipment

?

尝辞迟丑补谤…

?

What test gear did you use for the power sweep and what were the test points? I have a couple here I’d like to compare.

?

Tom, N8ZM

?

From: [email protected] <[email protected]> On Behalf Of Lothar baier
Sent: Monday, March 21, 2022 5:25 PM
To: [email protected]
Subject: Re: [HP-Agilent-Keysight-equipment] Protecting Equipment

?

Years ago I had this discussion with a PhD student ,? his argument was that the limiter doesn’t matter unless it is triggered so I set up a experiment doing a power sweep of the limiter and capturing the output spectrum for each point ,? he was quite surprised

?

From: [email protected] <[email protected]> On Behalf Of si_emi_01 via groups.io
Sent: Monday, March 21, 2022 4:16 PM
To: [email protected]
Subject: Re: [HP-Agilent-Keysight-equipment] Protecting Equipment

?

Lothar,

?

Correct. The same thing happens with Zeners below 6.2V and TVS Diodes (way too much Capacitance for RF Applications).

?

Look at the Breakover point of the RF Limiter (or TVS), it is a curve. Its Impedance changes below the breakover voltage as current flows through it and approaches the knee. If it is not specified, it can be obtained via a Curve Tracer or manually generated I-V Curves that you can produce with a simple Power Supply UUT and Resistance Standard or Resistance Decade.

?

Many Signal Integrity issues develop when they are not used properly and the curve is well understood.

?

Ross

?

From: [email protected] <[email protected]> On Behalf Of Lothar baier
Sent: Monday, March 21, 2022 2:28 PM
To: [email protected]
Subject: Re: [HP-Agilent-Keysight-equipment] Protecting Equipment

?

What many people seem to forget is the fact that the non linear behavior of a limiter kicks in way before the limiter is actually “triggered” ?although the distortions are not as bad

?

From: [email protected] <[email protected]> On Behalf Of si_emi_01 via groups.io
Sent: Monday, March 21, 2022 2:50 PM
To: [email protected]
Subject: Re: [HP-Agilent-Keysight-equipment] Protecting Equipment

?

Tom,

?

That is also why MIL-STD-461 CE102 requires a 20dB Pad on the signal path from the LISN to the Receiver. They do not recommend a Limiter due to non-linearities that Lothar points out.

?

Limiters can really hurt the Noise Figure of a receiver input at well.

?

Ross

?

From: [email protected] <[email protected]> On Behalf Of n8zmTWH@...
Sent: Sunday, March 20, 2022 7:22 PM
To: [email protected]
Subject: Re: [HP-Agilent-Keysight-equipment] Protecting Equipment

?

Limiters are very commonly recommended for Conducted Immunity testing in EMC labs where the Spec An input comes from the Line Impedance Stabilization Network (LISN) because the turn on/off transients can be quite high. The levels expected to be measured typically should be much less than the distortion level of the limiter. The recommended limiter in this case actually has an intentional 10 dB insertion loss.

?

Tom, N8ZM

?

From: [email protected] <[email protected]> On Behalf Of Lothar baier
Sent: Sunday, March 20, 2022 9:03 PM
To: [email protected]
Subject: Re: [HP-Agilent-Keysight-equipment] Protecting Equipment

?

A few words about limiters ,

Generally limiters are used in systems to protect receivers usually in RADAR or EW applications ,? in test systems the use of limiters s usually frowned upon ( with very few exceptions)? and for good reason !

First of there is a cost penalty that is not insignificant , the cost for a connectorized limiter usually starts at about $600 and can go up into the $2K range .

Limiters also degrade VSWR and add additional system loss .

But the by far biggest problem is that limiters are non linear devices and as such excellent harmonics generators , harmonics will start to occur before the limiter hits the hard threshold and worsens as the power increases so if you put one in front a spectrum analyzer you have to be cautious not to end up with tst results that are worse than they actually are

?

From: [email protected] <[email protected]> On Behalf Of Jim Ford via groups.io
Sent: Sunday, March 20, 2022 2:53 PM
To: [email protected]
Subject: Re: [HP-Agilent-Keysight-equipment] Protecting Equipment

?

Mini-Circuits sells limiters as well as DC blocks, but they only go up to 8 GHz.? Not sure about the price, but they are most likely competitive.? Assuming you actually need one.

?

Jim Ford

?

------ Original Message ------

From: "Matt Huszagh" <huszaghmatt@...>

To: [email protected]

Sent: 3/20/2022 6:52:14 AM

Subject: Re: [HP-Agilent-Keysight-equipment] Protecting Equipment

?

It's worth adding that a DC block only provides some measure of safety. The ramp up voltage of a DC signal is an AC signal and can pass right through your series cap. If you look at DC blocking filter designs like the HP 3048A option K23, there's a limiter after the series cap. It seems like the "right" devices for this sort of thing are RF power limiters. Unfortunately, the cost of many of these devices new from reputable manufacturers probably matches or exceeds what many people paid for their second-hand spectrum analyzer. Maybe there are good second-hand market options, but I haven't looked into to it. If you want something to reliably protect a low frequency SA, that shouldn't be too hard to design and build. But, if you want to get the full 22 GHz range out of your 8566, I expect that wouldn't be so easy.

?

Agilent E8257D 20GHz Signal generator ?, N9030A 50GHz spectrum analyzer , power sweep in 0.5Db steps from 0 to 30dBm CW

?

?

尝辞迟丑补谤…

?

What test gear did you use for the power sweep and what were the test points? I have a couple here I’d like to compare.

?

?

?

Years ago I had this discussion with a PhD student ,? his argument was that the limiter doesn’t matter unless it is triggered so I set up a experiment doing a power sweep of the limiter and capturing the output spectrum for each point ,? he was quite surprised

?

?

Lothar,

?

Correct. The same thing happens with Zeners below 6.2V and TVS Diodes (way too much Capacitance for RF Applications).

?

Look at the Breakover point of the RF Limiter (or TVS), it is a curve. Its Impedance changes below the breakover voltage as current flows through it and approaches the knee. If it is not specified, it can be obtained via a Curve Tracer or manually generated I-V Curves that you can produce with a simple Power Supply UUT and Resistance Standard or Resistance Decade.

?

Many Signal Integrity issues develop when they are not used properly and the curve is well understood.

?

Ross

?

?

What many people seem to forget is the fact that the non linear behavior of a limiter kicks in way before the limiter is actually “triggered” ?although the distortions are not as bad

?

?

Tom,

?

That is also why MIL-STD-461 CE102 requires a 20dB Pad on the signal path from the LISN to the Receiver. They do not recommend a Limiter due to non-linearities that Lothar points out.

?

Limiters can really hurt the Noise Figure of a receiver input at well.

?

Ross

?

?

Limiters are very commonly recommended for Conducted Immunity testing in EMC labs where the Spec An input comes from the Line Impedance Stabilization Network (LISN) because the turn on/off transients can be quite high. The levels expected to be measured typically should be much less than the distortion level of the limiter. The recommended limiter in this case actually has an intentional 10 dB insertion loss.

?

?

?

A few words about limiters ,

Generally limiters are used in systems to protect receivers usually in RADAR or EW applications ,? in test systems the use of limiters s usually frowned upon ( with very few exceptions)? and for good reason !

First of there is a cost penalty that is not insignificant , the cost for a connectorized limiter usually starts at about $600 and can go up into the $2K range .

Limiters also degrade VSWR and add additional system loss .

But the by far biggest problem is that limiters are non linear devices and as such excellent harmonics generators , harmonics will start to occur before the limiter hits the hard threshold and worsens as the power increases so if you put one in front a spectrum analyzer you have to be cautious not to end up with tst results that are worse than they actually are

?

?

?

尝辞迟丑补谤…

?

What test gear did you use for the power sweep and what were the test points? I have a couple here I’d like to compare.

?

?

?

Years ago I had this discussion with a PhD student ,? his argument was that the limiter doesn’t matter unless it is triggered so I set up a experiment doing a power sweep of the limiter and capturing the output spectrum for each point ,? he was quite surprised

?

?

Lothar,

?

Correct. The same thing happens with Zeners below 6.2V and TVS Diodes (way too much Capacitance for RF Applications).

?

Look at the Breakover point of the RF Limiter (or TVS), it is a curve. Its Impedance changes below the breakover voltage as current flows through it and approaches the knee. If it is not specified, it can be obtained via a Curve Tracer or manually generated I-V Curves that you can produce with a simple Power Supply UUT and Resistance Standard or Resistance Decade.

?

Many Signal Integrity issues develop when they are not used properly and the curve is well understood.

?

Ross

?

?

What many people seem to forget is the fact that the non linear behavior of a limiter kicks in way before the limiter is actually “triggered” ?although the distortions are not as bad

?

?

Tom,

?

That is also why MIL-STD-461 CE102 requires a 20dB Pad on the signal path from the LISN to the Receiver. They do not recommend a Limiter due to non-linearities that Lothar points out.

?

Limiters can really hurt the Noise Figure of a receiver input at well.

?

Ross

?

?

Limiters are very commonly recommended for Conducted Immunity testing in EMC labs where the Spec An input comes from the Line Impedance Stabilization Network (LISN) because the turn on/off transients can be quite high. The levels expected to be measured typically should be much less than the distortion level of the limiter. The recommended limiter in this case actually has an intentional 10 dB insertion loss.

?

?

?

A few words about limiters ,

Generally limiters are used in systems to protect receivers usually in RADAR or EW applications ,? in test systems the use of limiters s usually frowned upon ( with very few exceptions)? and for good reason !

First of there is a cost penalty that is not insignificant , the cost for a connectorized limiter usually starts at about $600 and can go up into the $2K range .

Limiters also degrade VSWR and add additional system loss .

But the by far biggest problem is that limiters are non linear devices and as such excellent harmonics generators , harmonics will start to occur before the limiter hits the hard threshold and worsens as the power increases so if you put one in front a spectrum analyzer you have to be cautious not to end up with tst results that are worse than they actually are

?

?

?

Years ago I had this discussion with a PhD student ,? his argument was that the limiter doesn’t matter unless it is triggered so I set up a experiment doing a power sweep of the limiter and capturing the output spectrum for each point ,? he was quite surprised

?

?

Lothar,

?

Correct. The same thing happens with Zeners below 6.2V and TVS Diodes (way too much Capacitance for RF Applications).

?

Look at the Breakover point of the RF Limiter (or TVS), it is a curve. Its Impedance changes below the breakover voltage as current flows through it and approaches the knee. If it is not specified, it can be obtained via a Curve Tracer or manually generated I-V Curves that you can produce with a simple Power Supply UUT and Resistance Standard or Resistance Decade.

?

Many Signal Integrity issues develop when they are not used properly and the curve is well understood.

?

Ross

?

?

What many people seem to forget is the fact that the non linear behavior of a limiter kicks in way before the limiter is actually “triggered” ?although the distortions are not as bad

?

?

Tom,

?

That is also why MIL-STD-461 CE102 requires a 20dB Pad on the signal path from the LISN to the Receiver. They do not recommend a Limiter due to non-linearities that Lothar points out.

?

Limiters can really hurt the Noise Figure of a receiver input at well.

?

Ross

?

?

Limiters are very commonly recommended for Conducted Immunity testing in EMC labs where the Spec An input comes from the Line Impedance Stabilization Network (LISN) because the turn on/off transients can be quite high. The levels expected to be measured typically should be much less than the distortion level of the limiter. The recommended limiter in this case actually has an intentional 10 dB insertion loss.

?

?

?

A few words about limiters ,

Generally limiters are used in systems to protect receivers usually in RADAR or EW applications ,? in test systems the use of limiters s usually frowned upon ( with very few exceptions)? and for good reason !

First of there is a cost penalty that is not insignificant , the cost for a connectorized limiter usually starts at about $600 and can go up into the $2K range .

Limiters also degrade VSWR and add additional system loss .

But the by far biggest problem is that limiters are non linear devices and as such excellent harmonics generators , harmonics will start to occur before the limiter hits the hard threshold and worsens as the power increases so if you put one in front a spectrum analyzer you have to be cautious not to end up with tst results that are worse than they actually are

?

?

On Mar 21, 2022, at 18:28, Yves Tardif via groups.io <yves_tardif@...> wrote:

?I think you are right, there is one more adjustment on the front of the C version, and it is a PEAK FILTER without any doubt for a YIG FILTER.

-----Message d'origine-----
De : [email protected] <[email protected]> De la part de Lothar baier
Envoyé : 21 mars 2022 19:15
? : [email protected]
Objet : Re: [HP-Agilent-Keysight-equipment] HP83595A/B (HP 8350) ROM firmware required

If I remember correctly than the main difference between the B and C was that it had a YIG filter added to produce a cleaner output signal

?

Hi David,

Thanks, I had seen that design on the ARF forum and was originally considering going that route but then had trouble finding the BFG35 transistors.? I was thinking that since the newer higher output MMIC's were readily available it might be a better/simpler design to go with the MMIC and reduce the parts count quite a bit.? Hopefully the RF choke and biasing wont complicate this too much for the lower frequencies.

Thanks
DB


"This design may help with avoiding re-inventing the wheel:

Hi,

I've read that it may be possible to use a MMIC amp to replace the failed hybrid output amps in the HP 8601a sweep generator.? I have some MiniCircuits GALI MMIC amps and wanted to try one of these.?

I am using AppCAD for the design calcs for blocking caps but have a question about the RF choke that is in series with the 200 ohm resistor on the bias power lead.
The MiniCircuits data says the impedance of the choke should be ~500ohms at the minimum operating freq (which is 100khz for the HP 8601a).? This is giving me a value of 795uH for 500ohm impedance.

The other limitation for the choke is the SRF self resonant freq.? The chokes that I'm finding in the 795uH range all seem to have SRF much lower than the max 110Mhz freq of the 8601a.? (most of the 800uH chokes I've found have SRF around 20-50Mhz).

I would appreciate any assistance on selecting the proper choke value that will be adequate for 100khz to 110Mhz.

Thanks
DB

The GALI-74 doesn’t produce enough power , the spec is 20dBm but since this is P1Db you need to be at least at 21dBm , the GALI-84 would work , I looked earlier and found the CMA-84 , not sure why the GALI-84 didn’t show in my search ,? the MAR-8 seems to make a good driver .

It will take me a day to download the models for the GALI and MAR and run a few simuations .

What you have to watch out for on gain blocks is that they usually have a lot of excess gain at low frequencies which can get you in trouble with stability , stability can also be impacted improperly designed bias networks as well and improper grounding ( need a lot of vias under the part to keep your ground inductance low ) most of the time companies like mini circuits don’t provide S-Parameters below 10MHz which makes assessing the gain at low frequencies without a simulator even harder !