RE: [TangerineSDR] Notes from PSWS / TangerineSDR call of 07-26-2021

[Dr. Nathaniel A. Frissell Ph.D.]

Thank you, Tom.

 

The recording will be available later today at https://youtu.be/AammohuygMw and hamsci.org/telecons.

 

73 de Nathaniel W2NAF

 

From: TangerineSDR <tangerines...@lists.tapr.org> On Behalf Of Tom McDermott via TangerineSDR
Sent: Monday, July 26, 2021 10:02 PM
To: TAPR TangerineSDR Modular Software Defined Radio <tanger...@lists.tapr.org>
Cc: Tom McDermott <tom....@gmail.com>
Subject: [TangerineSDR] Notes from PSWS / TangerineSDR call of 07-26-2021

 

 

Notes from PSWS / TangerineSDR call of 07-26-2021

 

1. Bill is using chart.js for magnetometer charting. He is setting up a database using Django web and database framework for Python.

 

2. Scotty is looking at the Intel (Altera) Arria 10 GX FPGA 10GX270 for the version 2 Data Engine (supporting 10GE). These FPGAs appear to be more available than the MAX10 FPGAs. The intention is to develop DE Ver 1 and DE Ver 2 in parallel while awaiting FPGA component availability. The 10 GX development boards are pretty expensive. 

 

-- Tom, N5EG

 

 

 

[Jonathan]

Hi Everyone,

Before the meeting ended last night, I wanted to discuss the issues I’ve experienced with my VLF active antenna. As I mentioned, the DC-DC converter (XP Power IP2415S) failed after I applied power. I turned the volume up because I wasn’t hearing sferics, and shortly after, I heard the squeal of the DC-DC converter. When I pulled out the smaller foam core partially, I heard a squeal coming from the DC-DC converter itself. 

I had a failure of the DC-DC converter when I first built the VLF preamp. It was being powered by the Raspberry Pi box (same power circuit as my text box). I had a power failure and when the power was restored, the DC-DC converter failed.  In both situations, it was being supplied with 30VDC, with the limit is 36V. The reason for using 30V is to overcome the voltage drop from long cable lengths.

After both failures, there was a low impedance across the input of the DC-DC converter causing the supply to be pulled down. Because the input and output are isolated, I figured the fault had to have been caused by a voltage spike on the input due to power being turned on and off. The datasheet mentions that it can withstand 50V spikes at 100ms, but I can’t think of anything else that could cause the DC-DC to fail, especially that 30V is used to power it and any voltage spike would ride on top of that 30V. I contacted XP Power about the issue. 

Here is the schematic of the power path. The blocks are the IP2415S DC-DC converter.

I’m powering the DC-DC converter with 30V from another DC-DC converter in the test box and Raspberry Pi box (Raspberry Pi box not shown, but same circuit). Keep in mind, the length is still relatively short in my testing, so cable inductance isn’t the primary reason for the spike. Plus, at long lengths, the voltage would be much lower due to voltage drop across the length, and I think this was why Paul might not have seen this issue when he originally tried and built this design. 

To mitigate this problem with the supply voltage at 30V, I decided to use a TVS diode across the input of the DC-DC converter. Here is the schematic: (my apologies for the TVS diode being backwards)

I chose a unidirectional TVS so there is also protection if the voltage spikes have any negative components. I picked a TVS with a working voltage of 33V, a Littel Fuse P4KE39A. According to the datasheet, it looks like this would be adequate. TVS diodes have a PN junction that is more rugged, with a larger surface area for greater current density. One thing I overlooked with this type of circuit protection is that once the TVS conducts, current from the power supply will be shunted too, known as the follow current, and in some situations, this may cause issues requiring a fuse. In this case, since the spikes are probably short duration, and since the DC-DC converter in the text/Pi box has short circuit protection, this should not be an issue. 

So the questions I have are: do you also think it could be voltage spikes that caused the failures, and if so, was my choice of TVS diode adequate. Also, one important question, if I connect the anode of the TVS diode to the receiver side ground, will I lose isolation? I know that during spike current conduction, I will. I’m sure I can still get adequate protection with the TVS diode anode connected to the negative of the isolated power loop too, but I just wanted to be sure. 

Lastly, since the magnetometer board will undergo another revision, I feel it may be worth it to add some sort or adjustable regulator or use of an external power source (not connected to the 5V pin of the Pi header) because even with paralleling wires or pairs, some installations will require very long lengths (especially since interference-free locations are often far away from the shack) and the voltage may drop to where the LDO cant regulate. Using a higher voltage at the Pi end (like I do with the VLF preamp) may be required for some installations. I think this should be considered if it already wasn’t addressed. 

Jonathan

KC3EEY

--
Please follow the HamSCI Community Participation Guidelines at http://hamsci.org/hamsci-community-participation-guidelines.
---
You received this message because you are subscribed to the Google Groups "HamSCI" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hamsci+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/SA0PR03MB5547CBF733EFEB4D829E659FF2E99%40SA0PR03MB5547.namprd03.prod.outlook.com.

[Khan Tran]

Hi Jonathan,

Cool update. I’m placing an order for a vlf receiver some time this week. 

Couldn’t attend meeting last night because I was busy but I’ll be there on Monday.

73 khan KE8QWB 

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/E010F57A-1232-47F1-BFC9-BDC41F1587E9%40gmail.com.

[John Ackermann N8UR]

Hi Jonathan --

Just a suggestion re fuses... look into PTC resettable fuses which are
available in small SMT as well as through-hole designs. Their advantage
is that after the overload is removed and they cool down, the circuit
comes back to life. That could be particularly handy at the far end of
a long wire. :-)

You need to do some work to figure out the right "hold" and "trip"
values to use, as well as the response time to make sure it can carry
the static current and still respond to the overload quickly enough.

73,
John

On 7/27/21 9:33 AM, Jonathan wrote:
> Hi Everyone,
>
> Before the meeting ended last night, I wanted to discuss the issues I’ve
> experienced with my VLF active antenna. As I mentioned, the DC-DC
> converter (XP Power IP2415S) failed after I applied power. I turned the
> volume up because I wasn’t hearing sferics, and shortly after, I heard
> the squeal of the DC-DC converter. When I pulled out the smaller foam
> core partially, I heard a squeal coming from the DC-DC converter itself.

> image3.jpeg
>
> I had a failure of the DC-DC converter when I first built the VLF
> preamp. It was being powered by the Raspberry Pi box (same power circuit
> as my text box). I had a power failure and when the power was restored,
> the DC-DC converter failed.  In both situations, it was being supplied
> with 30VDC, with the limit is 36V. The reason for using 30V is to
> overcome the voltage drop from long cable lengths.
>
> After both failures, there was a low impedance across the input of the
> DC-DC converter causing the supply to be pulled down. Because the input
> and output are isolated, I figured the fault had to have been caused by
> a voltage spike on the input due to power being turned on and off. The
> datasheet mentions that it can withstand 50V spikes at 100ms, but I
> can’t think of anything else that could cause the DC-DC to fail,
> especially that 30V is used to power it and any voltage spike would ride
> on top of that 30V. I contacted XP Power about the issue.
>
> Here is the schematic of the power path. The blocks are the IP2415S
> DC-DC converter.

> image1.jpeg
> I’m powering the DC-DC converter with 30V from another DC-DC converter
> in the test box and Raspberry Pi box (Raspberry Pi box not shown, but
> same circuit). Keep in mind, the length is still relatively short in my
> testing, so cable inductance isn’t the primary reason for the spike.
> Plus, at long lengths, the voltage would be much lower due to voltage
> drop across the length, and I think this was why Paul might not have
> seen this issue when he originally tried and built this design.
>
> To mitigate this problem with the supply voltage at 30V, I decided to
> use a TVS diode across the input of the DC-DC converter. Here is the
> schematic: (my apologies for the TVS diode being backwards)

> <mailto:nathaniel...@scranton.edu>> wrote:
>
>> Thank you, Tom.
>>
>> The recording will be available later today at

>> https://youtu.be/AammohuygMw <https://youtu.be/AammohuygMw> and
>> hamsci.org/telecons <http://hamsci.org/telecons>.
>>
>> 73 de Nathaniel W2NAF
>>
>> *From:* TangerineSDR <tangerines...@lists.tapr.org
>> <mailto:tangerines...@lists.tapr.org>> *On Behalf Of *Tom
>> McDermott via TangerineSDR
>> *Sent:* Monday, July 26, 2021 10:02 PM
>> *To:* TAPR TangerineSDR Modular Software Defined Radio
>> <tanger...@lists.tapr.org <mailto:tanger...@lists.tapr.org>>
>> *Cc:* Tom McDermott <tom....@gmail.com <mailto:tom....@gmail.com>>
>> *Subject:* [TangerineSDR] Notes from PSWS / TangerineSDR call of

>> 07-26-2021
>>
>> Notes from PSWS / TangerineSDR call of 07-26-2021
>>
>> 1. Bill is using chart.js for magnetometer charting. He is setting up
>> a database using Django web and database framework for Python.
>>
>> 2. Scotty is looking at the Intel (Altera) Arria 10 GX FPGA 10GX270
>> for the version 2 Data Engine (supporting 10GE). These FPGAs appear to
>> be more available than the MAX10 FPGAs. The intention is to develop DE
>> Ver 1 and DE Ver 2 in parallel while awaiting FPGA component
>> availability. The 10 GX development boards are pretty expensive.
>>
>> -- Tom, N5EG
>>
>> --
>> Please follow the HamSCI Community Participation Guidelines at
>> http://hamsci.org/hamsci-community-participation-guidelines

>> <http://hamsci.org/hamsci-community-participation-guidelines>.

>> ---
>> You received this message because you are subscribed to the Google
>> Groups "HamSCI" group.
>> To unsubscribe from this group and stop receiving emails from it, send
>> an email to hamsci+un...@googlegroups.com

>> <mailto:hamsci+un...@googlegroups.com>.

>> To view this discussion on the web visit
>> https://groups.google.com/d/msgid/hamsci/SA0PR03MB5547CBF733EFEB4D829E659FF2E99%40SA0PR03MB5547.namprd03.prod.outlook.com

>> <https://groups.google.com/d/msgid/hamsci/SA0PR03MB5547CBF733EFEB4D829E659FF2E99%40SA0PR03MB5547.namprd03.prod.outlook.com?utm_medium=email&utm_source=footer>.

>
> --
> Please follow the HamSCI Community Participation Guidelines at
> http://hamsci.org/hamsci-community-participation-guidelines

> <http://hamsci.org/hamsci-community-participation-guidelines>.

> ---
> You received this message because you are subscribed to the Google
> Groups "HamSCI" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to hamsci+un...@googlegroups.com

> <mailto:hamsci+un...@googlegroups.com>.

> To view this discussion on the web visit

> https://groups.google.com/d/msgid/hamsci/E010F57A-1232-47F1-BFC9-BDC41F1587E9%40gmail.com
> <https://groups.google.com/d/msgid/hamsci/E010F57A-1232-47F1-BFC9-BDC41F1587E9%40gmail.com?utm_medium=email&utm_source=footer>.

[Tom McDermott]

Hi Jonathan - your schematics do not show any bypass capacitors right next to the DC-DC converter input.

Normally with a long feed line there is considerable series inductance which might cause instability in such a converter.

Sometimes that can result in converter failure, other times perhaps an inability to regulate well.

Does the DC-DC converter datasheet or manual discuss input bypassing needs?

-- Tom, N5EG

Please follow the HamSCI Community Participation Guidelines at http://hamsci.org/hamsci-community-participation-guidelines.

---
You received this message because you are subscribed to the Google Groups "HamSCI" group.

To unsubscribe from this group and stop receiving emails from it, send an email to hamsci+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/778babd8-9595-528e-997f-b70322aa5c1f%40febo.com.

[Julius Madey]

Tom, agreed, I just sent Jonathan a note to that effect with the suggestion also that the series polarity protection diode be replaced by a protection diode in parallel with the TVS and perhaps a 10-20uF 50 volt aluminum electrolytic and a poly fuse in series which can open, either on overload or accidentally reversed polarity ..
73,
Jules-K2KGJ

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/CACO3nRSrLuf8ZC0cE3DWyohrV0UissOnSW_i%3D7hi4q63%2BU3E9A%40mail.gmail.com.

[Jonathan]

John, 

Thanks for the suggestion! I think I’ll start with the overvoltage age spike of 50V for 100ms or less based on the datasheet spec of the DC-DC converter. I was thinking of some sort of fusable circuit protection now that I’ll be using a TVS diode to mitigate stress from the follow current. 

Tom,

The datasheet wasn’t clear on the need for input filtering. On the output of both the IP2415S, I have 0.1uF ceramic capacitors. In the datasheet, it only mentions a spec of “Input Reflected Rated Current” with a series inductor and shunt capacitor. I don’t really understand exactly what this is, but it lists “20mAp-p through a 12uH inductor and 47uF capacitor”. Nothing else is mentioned about input protection or filtering.

In both situations that the DC-DC failed, the cable length was only 4’ of cat 5. This circuit that was proved and tested by Paul, including the receiver and schematic I attached, did not have such an issue, but the only difference is that the voltage was lower than 30V. That’s why I was thinking it was a turn-on/off spike. I think that in a lot of typical applications, the input voltage isn’t that high, so spikes don’t normally present a problem. The DC-DC that feeds the receiver DC-DC has never failed and it was always fed with 18V or less. 

Jules,

You recommend the Schottky diode in parallel with the TVS, then a series PTC, so during a spike, the TVS will conduct and the PTC would fuse, and during a reverse polarity condition, the Schottky diode will conduct with limited current due to the PTC fusing? I can see the benefit that the series PTC in normal a normal circuit condition would drop much less voltage that the Schottky diode. 

All in all, I agree with adding some capacitance to filter voltage spikes and reduce or eliminate them. I’ll need to be sure not to exceed the DC-DC max load capacitance, which is 47uF in this case. 

But, to me, the most plausible cause for the failures are turn on/off voltage spikes, even with a short length of cable. Adding electrolytics is a bit difficult due to the tight space constraint as seen in my previous email, so I’ll try out the TVS and a PTC. Other than voltage spikes, does anything else come to mind? 

Thanks guys. 

Jonathan

KC3EEY

On Tue, Jul 27, 2021 at 9:34 AM Jonathan <emum...@gmail.com> wrote:

[Tom McDermott]

Hi Jonathan - a concern is that the DC-DC converter input stage would become unstable and oscillate

with the cable inductance.  If that happens then there could be high frequency AC voltages present at the

input of the converter.  Would those over-voltage the input?  Possibly.   Would it radiate RFI from the cable?  Likely.

Good design dictates sufficiently low source impedance to the DC-DC converter.  Almost all linear regulators

require it.  An electrolytic and ceramic bypass pair right at the DC-DC input would be considered good design by many.

-- Tom, N5EG

[Phil Erickson]

Hi Tom,

  Isn't it also the case that conducted EMI from the DC-DC converter (not radiated) can be a large problem without proper bypass?  There is more than one way to corrupt the RF you want to measure.

73

Phil W1PJE

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/CACO3nRRfrsP3YdqOPRGpvgtF%3DQoXxBddrud3a%2BTg3Cf68%2B0pLA%40mail.gmail.com.

--

----
Phil Erickson
phil.e...@gmail.com

[Jonathan]

Hi Tom,

That makes a lot of sense. I’ll try fit a 0.1uF and a 22-47uF electrolytic on the feedline loop. 

Thanks again! 

Jonathan

KC3EEY

On Jul 28, 2021, at 10:55 AM, Tom McDermott <tom....@gmail.com> wrote:

Hi Jonathan - a concern is that the DC-DC converter input stage would become unstable and oscillate

with the cable inductance.  If that happens then there could be high frequency AC voltages present at the

input of the converter.  Would those over-voltage the input?  Possibly.   Would it radiate RFI from the cable?  Likely.

Good design dictates sufficiently low source impedance to the DC-DC converter.  Almost all linear regulators

require it.  An electrolytic and ceramic bypass pair right at the DC-DC input would be considered good design by many.

-- Tom, N5EG

On Wed, Jul 28, 2021 at 7:30 AM Jonathan <emum...@gmail.com> wrote:

John, 

Thanks for the suggestion! I think I’ll start with the overvoltage age spike of 50V for 100ms or less based on the datasheet spec of the DC-DC converter. I was thinking of some sort of fusable circuit protection now that I’ll be using a TVS diode to mitigate stress from the follow current. 

Tom,

The datasheet wasn’t clear on the need for input filtering. On the output of both the IP2415S, I have 0.1uF ceramic capacitors. In the datasheet, it only mentions a spec of “Input Reflected Rated Current” with a series inductor and shunt capacitor. I don’t really understand exactly what this is, but it lists “20mAp-p through a 12uH inductor and 47uF capacitor”. Nothing else is mentioned about input protection or filtering.

In both situations that the DC-DC failed, the cable length was only 4’ of cat 5. This circuit that was proved and tested by Paul, including the receiver and schematic I attached, did not have such an issue, but the only difference is that the voltage was lower than 30V. That’s why I was thinking it was a turn-on/off spike. I think that in a lot of typical applications, the input voltage isn’t that high, so spikes don’t normally present a problem. The DC-DC that feeds the receiver DC-DC has never failed and it was always fed with 18V or less. 

Jules,

You recommend the Schottky diode in parallel with the TVS, then a series PTC, so during a spike, the TVS will conduct and the PTC would fuse, and during a reverse polarity condition, the Schottky diode will conduct with limited current due to the PTC fusing? I can see the benefit that the series PTC in normal a normal circuit condition would drop much less voltage that the Schottky diode. 

All in all, I agree with adding some capacitance to filter voltage spikes and reduce or eliminate them. I’ll need to be sure not to exceed the DC-DC max load capacitance, which is 47uF in this case. 

But, to me, the most plausible cause for the failures are turn on/off voltage spikes, even with a short length of cable. Adding electrolytics is a bit difficult due to the tight space constraint as seen in my previous email, so I’ll try out the TVS and a PTC. Other than voltage spikes, does anything else come to mind? 

Thanks guys. 

Jonathan

KC3EEY

On Tue, Jul 27, 2021 at 9:34 AM Jonathan <emum...@gmail.com> wrote:

Hi Everyone,

Before the meeting ended last night, I wanted to discuss the issues I’ve experienced with my VLF active antenna. As I mentioned, the DC-DC converter (XP Power IP2415S) failed after I applied power. I turned the volume up because I wasn’t hearing sferics, and shortly after, I heard the squeal of the DC-DC converter. When I pulled out the smaller foam core partially, I heard a squeal coming from the DC-DC converter itself. 

[David G. McGaw]

Those sound like good values.  For a switcher it is critical to properly bypass the input, both for RF and for low frequencies.   As its current draw increases with lower voltage, it exhibits negative resistance and will oscillate given a high source impedance such as a long cable.

David N1HAC

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/59FAEE9C-11F5-4CDF-B960-2E583E42E89C%40gmail.com.

[Ward Silver]

A quick note - rectifiers, LEDs, or any non-linear devices connected to unshielded cables of any significant length will happily act as mixers or harmonic generators.  Whenever one of these is used, put a 0.01uF disc ceramic across it to bypass it at RF.

73, Ward N0AX

Virus-free. www.avast.com

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/59FAEE9C-11F5-4CDF-B960-2E583E42E89C%40gmail.com.

Virus-free. www.avast.com

[Julius Madey]

On that note, I wonder if it would be a good idea to put a common mode choke (multiple turns of the power feed line through an appropriate ferrite mix)  in the line ?  K9YC's audio RFI paper is a good reference.
73,
Jules-K2KGJ

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/CAFr7d%3DqbQQj4-iEcb_wFnPzbeG6XYO1Q6RV4RhTpX0ihyFxvEQ%40mail.gmail.com.

[Ward Silver]

Blocking common-mode RF current using ferrite chokes is a good technique.  So is using shielded cable for everything and metal enclosures bonded together.  Jim has several tutorials about this available on his web page (k9yc.com/publish.htm) - see the slide show about reducing received noise.  This was an evolution of a presentation on common-mode chokes (http://www.yccc.org/Articles/W1HIS/CommonModeChokesW1HIS2006Apr06.pdf) by Chuck Counselman, W1HIS in the context of reduced received noise. (The presentation was created before Type 31 ferrite became available which is a much better material for EMI suppression.)

73, Ward N0AX

Virus-free. www.avast.com

[HL Serra]

Good, Ward. I suggested this to one of the other Tangerine magnetometer guys. From hard experience, Jim K9YC's common mode chokes are good, but with long runs in high common mode EMI environments I have had to split the coax into smaller sections each with their own common mode chokes. Sometimes that is the only effective answer.

Cheers, and 73, Larry N6NC

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/CAFr7d%3DrPHez0M4YR68wUkCCTx7nEWhy5%2BFMZrUnMEgcp4xVXXQ%40mail.gmail.com.

[Julius Madey]

The large Fair Rite 31 material snapit split cores can be placed anywhere along a run of cable.  The 1" ID FairRite 0431177081 ($14.84 at Mouser) will take 4 turns of RG8/U and the 1.4" ID 2631181381 ($21.02 at Mouser) split core will take even more or larger diameter cable .... the 1.4" ID core does not have a plastic housing and has to be taped or tie wrapped to hold the halves together ... it's heavy and should be protected against impacts.

Chokes can be placed randomly along a longer cable run to avoid common resonance lengths. 

73,
Jules-K2KGJ

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/CAM4%3DgM92S_eb5OF6ynuDcGZTrcUnyc4zVfLU-HXbGTjxYXYRgA%40mail.gmail.com.

[Jonathan]

Phil,

Thanks for letting me know. I’ll reach out to Earle for more information.

Ward,

Thank you for that reminder and the great design tip. The feedline I will be using is shielded CAT6. One pair carries DC power and the other carries VLF audio. I plan on grounding the shield too. 

Jules, 

Would mix 31 be appropriate for VLF? It looks like it’s recommended for a frequency range of 1-300MHz, so it that case, it should eliminate any of those high frequency transients from EFT? They can’t be installed in the conduit, but can be installed on the feedline coming into the Raspberry Pi box.

Thanks.

Jonathan

KC3EEY

[George Byrkit]

I don't have the URLs handy at the moment, but I suggest that you google for:
NPS-EC-07-002
NPS-EC-10-001

Jim Breakhall was on the distribution list for these. I suspect Ward Silver has the links, too. They were in an article recently, but I regret not having saved the URLs.

73,
George Byrkit, K9TRV

-----Original Message-----
From: ham...@googlegroups.com <ham...@googlegroups.com> On Behalf Of Jonathan
Sent: Thursday, July 29, 2021 1:24 PM
To: Ward Silver <hwar...@gmail.com>
Cc: Julius Madey <hil...@fairpoint.net>; hamsci <ham...@googlegroups.com>; Tom McDermott <tom....@gmail.com>; TAPR TangerineSDR Modular Software Defined Radio <tanger...@lists.tapr.org>
Subject: Re: [HamSCI] RE: [TangerineSDR] Notes from PSWS / TangerineSDR call of 07-26-2021

Phil,

Thanks for letting me know. I’ll reach out to Earle for more information.

Ward,

Thank you for that reminder and the great design tip. The feedline I will be using is shielded CAT6. One pair carries DC power and the other carries VLF audio. I plan on grounding the shield too.

Jules,

Would mix 31 be appropriate for VLF? It looks like it’s recommended for a frequency range of 1-300MHz, so it that case, it should eliminate any of those high frequency transients from EFT? They can’t be installed in the conduit, but can be installed on the feedline coming into the Raspberry Pi box.




Thanks.

Jonathan

KC3EEY

On Jul 28, 2021, at 4:46 PM, Ward Silver <hwar...@gmail.com <mailto:hwar...@gmail.com> > wrote:



Blocking common-mode RF current using ferrite chokes is a good technique. So is using shielded cable for everything and metal enclosures bonded together. Jim has several tutorials about this available on his web page (k9yc.com/publish.htm <http://k9yc.com/publish.htm> ) - see the slide show about reducing received noise. This was an evolution of a presentation on common-mode chokes (http://www.yccc.org/Articles/W1HIS/CommonModeChokesW1HIS2006Apr06.pdf) by Chuck Counselman, W1HIS in the context of reduced received noise. (The presentation was created before Type 31 ferrite became available which is a much better material for EMI suppression.)

73, Ward N0AX

<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon> Virus-free. www.avast.com <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link>

On Wed, Jul 28, 2021 at 3:32 PM Julius Madey <hil...@fairpoint.net <mailto:hil...@fairpoint.net> > wrote:


On that note, I wonder if it would be a good idea to put a common mode choke (multiple turns of the power feed line through an appropriate ferrite mix) in the line ? K9YC's audio RFI paper is a good reference.
73,
Jules-K2KGJ


On 7/28/2021 4:02 PM, Ward Silver wrote:


A quick note - rectifiers, LEDs, or any non-linear devices connected to unshielded cables of any significant length will happily act as mixers or harmonic generators. Whenever one of these is used, put a 0.01uF disc ceramic across it to bypass it at RF.

73, Ward N0AX

<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon> Virus-free. www.avast.com <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link>

On Wed, Jul 28, 2021 at 11:41 AM Jonathan <emum...@gmail.com <mailto:emum...@gmail.com> > wrote:


Hi Tom,

That makes a lot of sense. I’ll try fit a 0.1uF and a 22-47uF electrolytic on the feedline loop.

Thanks again!

Jonathan
KC3EEY

On Jul 28, 2021, at 10:55 AM, Tom McDermott <tom....@gmail.com <mailto:tom....@gmail.com> > wrote:



Hi Jonathan - a concern is that the DC-DC converter input stage would become unstable and oscillate
with the cable inductance. If that happens then there could be high frequency AC voltages present at the
input of the converter. Would those over-voltage the input? Possibly. Would it radiate RFI from the cable? Likely.

Good design dictates sufficiently low source impedance to the DC-DC converter. Almost all linear regulators
require it. An electrolytic and ceramic bypass pair right at the DC-DC input would be considered good design by many.

-- Tom, N5EG






On Wed, Jul 28, 2021 at 7:30 AM Jonathan <emum...@gmail.com <mailto:emum...@gmail.com> > wrote:


John,

Thanks for the suggestion! I think I’ll start with the overvoltage age spike of 50V for 100ms or less based on the datasheet spec of the DC-DC converter. I was thinking of some sort of fusable circuit protection now that I’ll be using a TVS diode to mitigate stress from the follow current.




Tom,

The datasheet wasn’t clear on the need for input filtering. On the output of both the IP2415S, I have 0.1uF ceramic capacitors. In the datasheet, it only mentions a spec of “Input Reflected Rated Current” with a series inductor and shunt capacitor. I don’t really understand exactly what this is, but it lists “20mAp-p through a 12uH inductor and 47uF capacitor”. Nothing else is mentioned about input protection or filtering.

In both situations that the DC-DC failed, the cable length was only 4’ of cat 5. This circuit that was proved and tested by Paul, including the receiver and schematic I attached, did not have such an issue, but the only difference is that the voltage was lower than 30V. That’s why I was thinking it was a turn-on/off spike. I think that in a lot of typical applications, the input voltage isn’t that high, so spikes don’t normally present a problem. The DC-DC that feeds the receiver DC-DC has never failed and it was always fed with 18V or less.




Jules,

You recommend the Schottky diode in parallel with the TVS, then a series PTC, so during a spike, the TVS will conduct and the PTC would fuse, and during a reverse polarity condition, the Schottky diode will conduct with limited current due to the PTC fusing? I can see the benefit that the series PTC in normal a normal circuit condition would drop much less voltage that the Schottky diode.




All in all, I agree with adding some capacitance to filter voltage spikes and reduce or eliminate them. I’ll need to be sure not to exceed the DC-DC max load capacitance, which is 47uF in this case.

But, to me, the most plausible cause for the failures are turn on/off voltage spikes, even with a short length of cable. Adding electrolytics is a bit difficult due to the tight space constraint as seen in my previous email, so I’ll try out the TVS and a PTC. Other than voltage spikes, does anything else come to mind?




Thanks guys.

Jonathan

KC3EEY

On Tue, Jul 27, 2021 at 9:34 AM Jonathan <emum...@gmail.com <mailto:emum...@gmail.com> > wrote:


Hi Everyone,


Before the meeting ended last night, I wanted to discuss the issues I’ve experienced with my VLF active antenna. As I mentioned, the DC-DC converter (XP Power IP2415S) failed after I applied power. I turned the volume up because I wasn’t hearing sferics, and shortly after, I heard the squeal of the DC-DC converter. When I pulled out the smaller foam core partially, I heard a squeal coming from the DC-DC converter itself.
<image3.jpeg>



I had a failure of the DC-DC converter when I first built the VLF preamp. It was being powered by the Raspberry Pi box (same power circuit as my text box). I had a power failure and when the power was restored, the DC-DC converter failed. In both situations, it was being supplied with 30VDC, with the limit is 36V. The reason for using 30V is to overcome the voltage drop from long cable lengths.


After both failures, there was a low impedance across the input of the DC-DC converter causing the supply to be pulled down. Because the input and output are isolated, I figured the fault had to have been caused by a voltage spike on the input due to power being turned on and off. The datasheet mentions that it can withstand 50V spikes at 100ms, but I can’t think of anything else that could cause the DC-DC to fail, especially that 30V is used to power it and any voltage spike would ride on top of that 30V. I contacted XP Power about the issue.


Here is the schematic of the power path. The blocks are the IP2415S DC-DC converter.

I’m powering the DC-DC converter with 30V from another DC-DC converter in the test box and Raspberry Pi box (Raspberry Pi box not shown, but same circuit). Keep in mind, the length is still relatively short in my testing, so cable inductance isn’t the primary reason for the spike. Plus, at long lengths, the voltage would be much lower due to voltage drop across the length, and I think this was why Paul might not have seen this issue when he originally tried and built this design.

To mitigate this problem with the supply voltage at 30V, I decided to use a TVS diode across the input of the DC-DC converter. Here is the schematic: (my apologies for the TVS diode being backwards)

I chose a unidirectional TVS so there is also protection if the voltage spikes have any negative components. I picked a TVS with a working voltage of 33V, a Littel Fuse P4KE39A. According to the datasheet, it looks like this would be adequate. TVS diodes have a PN junction that is more rugged, with a larger surface area for greater current density. One thing I overlooked with this type of circuit protection is that once the TVS conducts, current from the power supply will be shunted too, known as the follow current, and in some situations, this may cause issues requiring a fuse. In this case, since the spikes are probably short duration, and since the DC-DC converter in the text/Pi box has short circuit protection, this should not be an issue.

So the questions I have are: do you also think it could be voltage spikes that caused the failures, and if so, was my choice of TVS diode adequate. Also, one important question, if I connect the anode of the TVS diode to the receiver side ground, will I lose isolation? I know that during spike current conduction, I will. I’m sure I can still get adequate protection with the TVS diode anode connected to the negative of the isolated power loop too, but I just wanted to be sure.

Lastly, since the magnetometer board will undergo another revision, I feel it may be worth it to add some sort or adjustable regulator or use of an external power source (not connected to the 5V pin of the Pi header) because even with paralleling wires or pairs, some installations will require very long lengths (especially since interference-free locations are often far away from the shack) and the voltage may drop to where the LDO cant regulate. Using a higher voltage at the Pi end (like I do with the VLF preamp) may be required for some installations. I think this should be considered if it already wasn’t addressed.

Jonathan
KC3EEY





On Jul 27, 2021, at 7:21 AM, Dr. Nathaniel A. Frissell Ph.D. <nathaniel...@scranton.edu <mailto:nathaniel...@scranton.edu> > wrote:



















Thank you, Tom.







The recording will be available later today at

https://youtu.be/AammohuygMw <https://youtu.be/AammohuygMw> and hamsci.org/telecons <http://hamsci.org/telecons> .

73 de Nathaniel W2NAF









From: TangerineSDR <tangerines...@lists.tapr.org <mailto:tangerines...@lists.tapr.org> >

On Behalf Of Tom McDermott via TangerineSDR


Sent: Monday, July 26, 2021 10:02 PM

To: TAPR TangerineSDR Modular Software Defined Radio <tanger...@lists.tapr.org <mailto:tanger...@lists.tapr.org> >


Cc: Tom McDermott <tom....@gmail.com <mailto:tom....@gmail.com> >

Subject: [TangerineSDR] Notes from PSWS / TangerineSDR call of 07-26-2021























Notes from PSWS / TangerineSDR call of 07-26-2021















1. Bill is using chart.js for magnetometer charting. He is setting up a database using Django web and database framework for Python.















2. Scotty is looking at the Intel (Altera) Arria 10 GX FPGA 10GX270 for the version 2 Data Engine (supporting 10GE). These FPGAs appear to be more available than the MAX10 FPGAs. The intention is to develop DE Ver 1 and DE Ver 2 in parallel

while awaiting FPGA component availability. The 10 GX development boards are pretty expensive.















-- Tom, N5EG













































--


Please follow the HamSCI Community Participation Guidelines at http://hamsci.org/hamsci-community-participation-guidelines.


---


You received this message because you are subscribed to the Google Groups "HamSCI" group.

To unsubscribe from this group and stop receiving emails from it, send an email to hamsci+un...@googlegroups.com <mailto:hamsci+un...@googlegroups.com> .


To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/SA0PR03MB5547CBF733EFEB4D829E659FF2E99%40SA0PR03MB5547.namprd03.prod.outlook.com <https://groups.google.com/d/msgid/hamsci/SA0PR03MB5547CBF733EFEB4D829E659FF2E99%40SA0PR03MB5547.namprd03.prod.outlook.com?utm_medium=email&utm_source=footer> .

--
Please follow the HamSCI Community Participation Guidelines at http://hamsci.org/hamsci-community-participation-guidelines.
---
You received this message because you are subscribed to the Google Groups "HamSCI" group.

To unsubscribe from this group and stop receiving emails from it, send an email to hamsci+un...@googlegroups.com <mailto:hamsci+un...@googlegroups.com> .
To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/59FAEE9C-11F5-4CDF-B960-2E583E42E89C%40gmail.com <https://groups.google.com/d/msgid/hamsci/59FAEE9C-11F5-4CDF-B960-2E583E42E89C%40gmail.com?utm_medium=email&utm_source=footer> .



<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon> Virus-free. www.avast.com <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link>

--
Please follow the HamSCI Community Participation Guidelines at http://hamsci.org/hamsci-community-participation-guidelines.
---
You received this message because you are subscribed to the Google Groups "HamSCI" group.

To unsubscribe from this group and stop receiving emails from it, send an email to hamsci+un...@googlegroups.com <mailto:hamsci+un...@googlegroups.com> .
To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/CAFr7d%3DqbQQj4-iEcb_wFnPzbeG6XYO1Q6RV4RhTpX0ihyFxvEQ%40mail.gmail.com <https://groups.google.com/d/msgid/hamsci/CAFr7d%3DqbQQj4-iEcb_wFnPzbeG6XYO1Q6RV4RhTpX0ihyFxvEQ%40mail.gmail.com?utm_medium=email&utm_source=footer> .

--
Please follow the HamSCI Community Participation Guidelines at http://hamsci.org/hamsci-community-participation-guidelines.
---
You received this message because you are subscribed to the Google Groups "HamSCI" group.

To unsubscribe from this group and stop receiving emails from it, send an email to hamsci+un...@googlegroups.com <mailto:hamsci+un...@googlegroups.com> .
To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/29865318-FCCB-4A1E-B3DB-EFF738A19F04%40gmail.com <https://groups.google.com/d/msgid/hamsci/29865318-FCCB-4A1E-B3DB-EFF738A19F04%40gmail.com?utm_medium=email&utm_source=footer> .

[Julius Madey]

Jonathan, I think that mix 75 would probably be better for the VLF range and is available in a snapit core with 0.7inch ID  part no. 0475176451.  Also available in other sizes with solid core.  Read up on the references that Ward sent for more detailed information on application.
73,
Jules-K2KGJ

[Jonathan]

Jules,

Thank you for the part number! I’ll order a few. 

George,

Thank you for those documents!

All,

Thank you for all of your great advice, help, and assistance! This was a really good discussion!

Thanks again.

Jonathan

KC3EEY

<image1.jpeg>

I’m powering the DC-DC converter with 30V from another DC-DC converter in the test box and Raspberry Pi box (Raspberry Pi box not shown, but same circuit). Keep in mind, the length is still relatively short in my testing, so cable inductance isn’t the primary reason for the spike. Plus, at long lengths, the voltage would be much lower due to voltage drop across the length, and I think this was why Paul might not have seen this issue when he originally tried and built this design. 

To mitigate this problem with the supply voltage at 30V, I decided to use a TVS diode across the input of the DC-DC converter. Here is the schematic: (my apologies for the TVS diode being backwards)

[Dana Whitlow]

Guys,

The IP2415S DC-DC converter is clearly a switcher, which means that within its regulation

range, the input current is a decreasing function of the input voltage for a given load.  This

behavior is driven by fundamental "conservation of energy" considerations.

Thus, at low to moderate frequencies, the input has a negative resistance characteristic.

Depending on the design of an input LC filter in front of the converter, and on the impedance

of the raw source, the system can oscillate, which is probably not good for either the converter

or its load.

One common filter is a series inductor, followed by a shunt capacitor, followed by the

converter's input.  If the raw source impedance is low, then the AC impedance seen by

the converter's input is peaked around the resonant frequency of the filter's inductor and

capacitor, which will cause oscillation if that resonant frequency is in the right range.

So, in strange cases involving switching converters or regulators, perhaps the first 

thing to look for is oscillation of the voltage at the converter's input with an o'scope.  

Such problems can often be quelled by Q-spoiling the filter's resonance, albeit at the

sacrifice of some efficiency and/or filter effectiveness.

Nothing's ever simple!

Dana   K8YUM

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/CEED81D7-C092-4FC5-A974-58F00762DB6D%40gmail.com.

[Bob McGwier]

Reminder that CAT6 has twisted pairs inside the shield.  This already provides common mode suppression unless the source is in the near field of the wire.  For high power near field sources, the ferrite, etc is probably required.

Bob

N4HY

On Sat, Jul 31, 2021 at 8:52 AM Jonathan <emum...@gmail.com> wrote:

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/CEED81D7-C092-4FC5-A974-58F00762DB6D%40gmail.com.