TX Temperatures

[Brock Nanson]

I was surprised at just how small the final output FETs (I assume?) are on the HL2 board.  The fact that they manage heat with such a relatively simple heat sink is also interesting.

Putting the assembly together with the tight clearances and some difficulty getting thermal paste where I really wanted it, has made me very aware of temperature.  Using Thetis, I can see the temperature climb when TX is triggered.  This makes me wonder about a few things that someone might be able to answer..

1. How is the displayed temperature in Thetis measured?  I'm wondering how accurately it indicates the temperature of the FETs themselves.
2. What maximum temperature is considered safe for repeated heat/cool cycles?
3. Has anyone related the rate of temperature change to a properly installed heat sink vs. a poorly installed one?
4. Is there any point in adding the older revision copper heat sink to the top of the devices, bolted to the side of the case?  The split case makes this look like a tough install.

Depending on what the HL2 is doing on receive and I suppose, the ambient temperature in my shack, the HL2 seems to settle in around 30 - 31 C.  Seems reasonable to me.

Running WSJT-X in WSPR mode, I've tried 1 watt (according to Thetis) and also full output at around 5 watts.  Surprisingly, the two minute transmit cycles (well, less the decode period at the end) seem to bump the temperature up to similar maximums, despite the different power levels.

1 watt takes the temp up to about 44C.  With only a single two minute receive cycle to cool, another TX cycle bumps it to 46C.  Setting the power to 5 watts after a long period of RX only and the temperature hits 47C.  Going to TX on the next cycle (with only the few seconds of decode to cool) pushes it to just under 51C (50.7).

I feel like my heat sink must be doing fairly well, because 30 seconds after the 50.7C peak, the reading was back down to 38.3.  The whole box was warmer than it had been and 120 seconds later it was at 36.1.

So I'm wondering, has anyone decided what is safe and how rate of change of temperature relates to heat sink effectiveness?

73

Brock VA7AV

[Steve Haynal]

Hi Brock,

There is a temperature sensor Q6 in the corner of the board where the LDMOS final amplifier is. During development, I measured the temperature of the LDMOS device cases with a laser thermometer and compared with the measurement from the Q6. There is good correlation. What matters is the internal junction temperature of the LDMOS devices and the datasheet describes how to calculate this based on the external case temperature. After going through these calculations, I decided that 55C as measured by Q6 is a good conservative upper limit for operation. The gateware will turn of TX if the temperature hits 55C, so you have protection.

Your setup does appear to be running a little hot, but not enough to worry me. It is a bit hard to specify a normal temperature as there is variation in the sensor, in a user's build, and the ambient temperature and humidity. Did you install the heat shim? Did you sand away the coating in the PCB slot? These two make the biggest difference. I abandoned the top side heat sink as I could only measure about 1 degree difference with and without. You can also install an internal fan as described in the wiki, but people have had better results by using an external fan blowing on the HL2 with feet so that air can flow under and over.

Over 600 HL2 units have sold and I have yet to hear of the finals being destroyed by temperature. There have been a couple of cases where shorting caused final damage.

73,

Steve

kf7o

[Brock Nanson]

Hi Steve,

I'm thinking that SWR will have an influence here, so I tried documenting what I've found.  I'm wondering whether the band in use also has an effect?

I tried 80 m with an SWR of 1.4:1 and also 20 m with an SWR of 1.1:1.  The collected data and some quick graphs are at the link:

https://docs.google.com/spreadsheets/d/1ElQTJbcyhio80dutVvWYH6Om5BaLgsvhxHVyrphWV54/edit?usp=sharing

The 20 m test was comprised of two back to back WSPR transmissions at full power.  The drop between is the 6 second decoding pause before going to TX again.  I'd like to go back and continue the transmissions to see if the curve will flatten before the gateware shuts things down at 55 C, but I'm not sure I want to stress things too much!  The second TX climb is a little flatter than the first I think, but I don't see it really bending towards an equilibrium.

The 80 m test tab on the spreadsheet was a single WSPR transmission.  SWR is higher and Thetis tells me the output is only 4 Watts.  That might be a calibration thing, or maybe it just doesn't like 80 m as much.  I note that the temperature climbed higher during the first transmission than it did with the flat match on 20 m.

Is there enough heat sink in the design to allow continuous output at full power at a reasonable temperature?

The very sharp drop in temperature when TX ends suggests to me that my heat sink is working fairly well, but I have no frame of reference for this.  I did remove the anodizing in the rail area and also at the point of contact underneath where the metal wedge sits.  And of course, used some Arctic Silver in the rail and on both contact points of the wedge.  Sliding everything together pretty much scrapes it away though, so I don't know how effectively it's working.  I think I ended up cleaning more of it off my hands than ended up on the surfaces... LOL. 

Based on what you see in my graphs, do you think I'm advised to take everything apart and try again, or be happy with how it's doing?  I'm not worried about it cooking, now that I understand how the Gateware deals with this, but - you know - sometimes I can be too much of a perfectionist... ;-)

Brock

VA7AV

[Simon]

A pic of the over temp cure.. silent. Cheap. 40mm 12 v fan decoupled from supply.on all the time.. no shim fitted.no rf noise.

Just saying..silly easy  sorts all issues.even looks good.

Simon g0zen 

On 17 Apr 2021, at 23:31, Brock Nanson <br...@nanson.net> wrote:

Hi Steve,

--
You received this message because you are subscribed to the Google Groups "Hermes-Lite" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hermes-lite...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/9db8f869-6356-4bc6-9027-5ae62e060fedn%40googlegroups.com.

[Steve Haynal]

Hi Brock,

I wouldn't worry about it. In the past I've done 6 back-to-back WSPR transmit and all was fine. The temperature rise slows down as it gets higher and as seen if you extrapolate your data. Will the HL2 transmit continuously without overheating in all environments? No. Will the HL2 transmit ~50% of the time with overheating in normal environments? Yes.

73,

Steve

kf7o

[Brock Nanson]

Thanks Steve, I appreciate your comments.  A 50% duty cycle is realistically the worst I'd subject mine to in real world use... FT8 with only 15 seconds of TX is pretty trivial, temperature-wise.  WSPR would be much less than 50%, although the two minutes obviously takes the temperature higher.  The fact that the gateware has a failsafe built in is good enough for me though - I'll consider this question answered - thanks again!

Interesting what Simon did with his, and the fact that the fan is able to replace the heatsink shim.  It would be interesting to know what sort of duty cycle that configuration, but *with* the heatsink, could manage.  The heatsink is on the opposite side of the board from the fan though, so you'd have a combination of heat dissipation methods in use.  Also interesting to contemplate whether the fan would fool the temperature sensor into thinking the finals are cooler than they actually are...

73

Brock

VA7AV

[Simon]

Hi all

I feel that the temp monitoring is worrying too many people..

Look at from this point of view..

Yaesu ftdx 5000 or whatever.

Has no indication on screen of what temp is at...IT does not stop one using the rig on ft8.or rag chewing at 100w for hours on end... gets too hot fan comes on..

I think people are worrying over nothing..( caused by the fact that there is a temp monitoring facility.)

If one lives in the northern territories in Australia or on the equator then add a fan..its easy and WORKS. Its CHEAP.( £4..$5)  oh and silent too.( unless your hl2 is under your pillow at night.) 

Job done..maybe even add a thermister and transistor control if want to get flash.)

So simple ..And if one can not do above, maybe one should have bought an 7300 etc..HL2 is for curious experimenting types..( NOT SAYING YOU ARE IN THAT BRACKET!! please do not take offence..)

Best regards Simon g0zen 

On 19 Apr 2021, at 19:20, Brock Nanson <br...@nanson.net> wrote:

Thanks Steve, I appreciate your comments.  A 50% duty cycle is realistically the worst I'd subject mine to in real world use... FT8 with only 15 seconds of TX is pretty trivial, temperature-wise.  WSPR would be much less than 50%, although the two minutes obviously takes the temperature higher.  The fact that the gateware has a failsafe built in is good enough for me though - I'll consider this question answered - thanks again!

--

You received this message because you are subscribed to the Google Groups "Hermes-Lite" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hermes-lite...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/d858f566-8f35-447b-8d1f-e4380ba277a8n%40googlegroups.com.

[Brock Nanson]

Hi Simon,

I definitely understand your point.  I'm not worry about temperature as much as I'm interested in it.  :-)

The difference between HL2 and radios from big manufacturers is that the big guys have a team of well-paid RF engineers designing the boards, with all the lab equipment they could ask for.  They've built many different radios over the years and each model taught them something.  If one of those radio models burnt up from too much TX heat, there would be big trouble for the company and brand reputation.

HL2 is a different animal IMHO.  It's a work in progress, as witnessed by the existence of this group.  For that reason, my expectations are lower than for those expensive models from Yaesu, Kenwood, Icom, etc... (Hell, my expectations were so low, I was *elated* when I plugged the power in and LEDs started flashing - LOL!!) The temperature points I collected were mostly to satisfy my curiosity, with a little bit of wondering about how much variation was likely due to my own construction experiences and challenges.  And thinking about how to make improvements.  I see this thread as breadcrumbs for future HL2 owners who have similar questions.  And potentially seeds of improvement for future iterations of the radio...although I'm not sure there's much to improve.

I'm a curious person and worse yet, I'm also an engineer... which means my curiosity borders on pedanticism.... ;-)

FWIW, I'm quite shocked at just how good this radio (and the software out there to support it) actually is.  I was excited that it powered up and detected signals... the amazing performance I'm seeing after a period of use is a wonderful bonus and a phenomenal achievement in my view! 

Brock

[Simon]

Hi

No longer own this hl2, but i took power from 12v in, decoupled and added inductance etc to stop any noise from fan.

Its a tiny 40mm fan.no noise .

I never saw temp above 40c running it flat out with at time ambient temp 30c

Simon

On 18 Apr 2021, at 15:33, Marcos X <msp...@gmail.com> wrote:



Yes, just above the transistors, what temperatures does it maintain? Did you check it? where you connected the pins of 12V or less

El dom, 18 abr 2021 a las 15:18, Simon (<ohhelln...@gmail.com>) escribió:

A pic of the over temp cure.. silent. Cheap. 40mm 12 v fan decoupled from supply.on all the time.. no shim fitted.no rf noise.

<image0.jpeg>

Just saying..silly easy  sorts all issues.even looks good.

Simon g0zen 

To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/9db8f869-6356-4bc6-9027-5ae62e060fedn%40googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Hermes-Lite" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hermes-lite...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/5034A35F-7629-4A00-82CB-CF41B98E7B8D%40gmail.com.

[Steve Haynal]

Hi Brock,

I like your attitude. The HL2 is definitely more akin to a DIY/homebrew/kit radio versus a true commercial product, although it sits somewhere in the spectrum between those two extremes. Those interested in the HL2 should definitely expect a steeper learning curve and some rough areas in documentation and features. They should be comfortable with software, home networking and some light soldering and assembly. But those who accept this are often very happy with the HL2.

For a Hermes-Lite 3.0, I'd probably look into a better heat spreader. I consider the current heat dissipation adequate and meeting all specs, but it can run hotter than some people like.

73,

Steve

kf7o

[Pierre Martel]

Can I suggest that in a 3.0 HL2 that you give access to both the TX and RX ports after the filters. 

Since this radio is a homebrew kit people that use it are also playing with antennas. 

I for one would really like to have the possibility to plug a LOG (loop on the ground) for my receive and my 43 foot vertical for TX (with a tuner) so that I can have a low noise receiving antenna and a pretty good TX antenna. 

Right now the low level TX output is kind of unused by most HL2 users. unless I am wrong and feel free to put me back into the right tracks.

To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/cef7c4eb-1074-4568-a771-269104cc89a2n%40googlegroups.com.

[John Williams]

I am using the low power output for TX and a separate receive antenna. The amplifier I built has an extra relay on the T/R board that allows me to select the RX antenna. I can use the main antenna or an alternate beverage antenna. Having a separate RX port could allow this to occur using the onboard PA if the amp needed a higher drive value. Effectively what you need is to be able to disable the onboard T/R relay and manage that function outboard. 

You can see what I am doing on my QRZ.com page.

John W9JSW

To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/CAL8dJwpU1y27bA9PERi1SB%3D4YLtk8MRmNQhsXTwMfNVW6TP_8A%40mail.gmail.com.

[Matthew]

This is possible without a HL2 version 3.0. The HL2 PCB has separate TX and RX ports. The HL2-MRF101 PCB uses this to have the option to switch (in software) between a dedicated RX antenna and a TX antenna for receive.

What would be needed is a change to the N2ADR filter board. It just becomes a PCB layout logistics challenge. I think this is not too difficult to add a small surface mount relay for this (e.g. OMRON G6KU-2F).

73 Matthew M5EVT.

[Pierre Martel]

I know it can be done, Just adding a small relay controlled by a line from the filter. But that is not the point. 

I really think that those would be more useful then the low power output, Or in a new design of the filter, there could be a low power output (that come from the HL2 anyways, and a TX output that can be used as the RX feed but a third sma that would give a RX only input. I don't know how many control lines are still available out of the chip on the filter board.  But could it be done? 

To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/93c38cad-4d39-4b17-85ed-49540472605fn%40googlegroups.com.

[Brock Nanson]

Well, Steve... all I can say is "Now you've done it!"... mentioned "HL3"... LOL

This could easily become a new thread, but it would probably get very long very fast - I'm sure everyone has a wish list for the next generation design!

A few things that I'd like to see, with varying levels of effort and blue-sky dreaming involved...:

• A re-think of the enclosure design and I/O locations on the board.  A slightly bigger box would allow the ideas in the next bullet points to work.
• SO-239 connectors.  Chassis mounted for durability.  Am I allowed to say I hate, hate, hate SMA?  Well, I hate SMA.  LOL...  I use a short piece of coax with SMA on one end and SO-239 on the other to serve as a strain relief.  Too many connections and SMA durability isn't fantastic.
• Avoid connectors on both front and back of chassis.  If not enough room for everything out the back, then perhaps the side.  Ethernet and power out the front make for some awkward bends on the desk.  Key/PTT is logically out the front IMHO, about the only exception.
• An optional modular amplifier board, to fit in the larger chassis as an upgrade, like the N2ADR board does.  40 or 50 watts would satisfy me, but 90 or 100 would make it functionally comparable with the big commercial options out there, while retaining the tinkerer/homebrew appeal.
• An optional VHF/UHF (or higher) modular board(s).
• An optional ATU.
• A fuse, perhaps part of the amplifier 'upgrade'.
• A review of heat management, especially if an amplifier board is made available.  Maybe this means a fan option, but I'd prefer a completely passive solution - which a bigger chassis might allow with some heat sinks.
• Consider detaching the LEDs, 1/8th jack from the board and put on a separate board with ribbon cable, to allow it to be connected to a face plate without the need for the main board to be accurately located in the chassis.
• If a larger chassis is used, there is more ability to expand the board to allow for easier soldering of additional optional connections. I look at the rear plate with traces and think 'no chance in hell I'm ever going to solder anything to that without making a huge mess'... some obviously can, but I'm more comfortable with old-school board construction where I can use a soldering iron and not require a hot air system and jeweler's loupe.  Let Makerfabs do the finicky construction, but leave enough space on the board for the end users to do some of the additions that I understand are possible.
• It's interesting to contemplate how several HL2(3)s could be built into a 1U or 2U rackmount chassis, to allow a club to build a relatively inexpensive multiband remote station (for example).  Add some rackmount computers to support them and a data pipe and good to go.

None of these ideas make the product cheaper, I realize that.  Some are years away, if ever.  Perhaps HL2 could remain in production as a smaller, cheaper option, with an HL3 as a more expandable but costlier upgrade?  Most of my ideas really circle back to modularity and expandability.  Optional modules could be developed more easily by others, if the HL3 was built with more attention to easy expandability.

I also own a Flex 6400.  It's the same as the HL2, only very different...(!)  I could do a comparative review of both, but that would go on for pages.  Bottom line,  HL2 isn't nearly as inferior as one would think at first glance... especially if you factor in the price.  HL2 is the QRP equivalent of the 6400 in many ways - something I don't think FlexRadio would like to hear me say!  I have more fun with HL2 and I enjoy both the HL2 community and community support that is tied to the HL2 more than that of Flex.  I don't like that bugs are allowed to remain in SmartSDR for... (forever?)... because it's a commercial product with a significant price tag!  Software for the HL2?  Several options and based on the price, there should be exactly zero support - but clearly that's not the case and I love the efforts the developers are making to improve the different software options.  I can't say enough good things about those people who are doing the heavy lifting.  It feels much more like my experiences with Linux over the past 25 years, which have been overwhelmingly positive, even though I can't code to save my life.  

If could get 75 watts out of single box built around HL2(3), with robust construction to withstand typical Field Day conditions, I'd probably be looking to sell my Flex.  The smiles-per-dollar value of HL2 is just that good.

73,

Brock

VA7AV

[Heath Petty]

Brock,

Have you seen this video:

https://www.youtube.com/watch?v=V58GDuoRlfY

it goes into some depth of how Elecraft designed their K4 radio. There is no reason that the Hermes board couldn't used in a similar design as the SDR board/module in that radio. If you want compactness and portability, have a search for the piHPSDR project. I was going to build a piHPSDR with the hermes into an apache case from harbor freight, but I am having issues with pulseaudio crashing on my pi4. At the end of this month I should be receiving my hackboard 2 (celeron based SBC). Nice thing with this board is it runs off of 12v, so If I can find a 12v screen, I won't have to mess with any DC voltage converters. For more kick, I have purchased an XPA125B amplifier, and down the line I will look more into the MRF100 based 100W amp (or even build a nice 500W amp).

I guess my point is that the current HL2 board can easily be used as a component in a bigger design, and in the opensource spirit, if you have an itch, scratch it :D

-Heath

[ron.ni...@gmail.com]

Hi Brock,

You could add almost all the features you want with the current HL2 b9, just by mounting it on a carrier circuit board (maybe design your own?) inside a larger box (or rackmount case).  

I hate SO-239 because the cables are far heavier to carry around for POTA/SOTA portable ops.  I'm more interested in a smaller lighter plastic case, with maybe just a small fan heatsink near the power transistors.

73,

Ron

n6ywu

[Brock Nanson]

I'm going to watch the full Elecraft video this evening - it's too long to get through on a lunch break... ;-)

I agree with both Ron and Heath that we absolutely can do pretty much whatever we want today, with the current configuration.  I think what my comments boil down to is, the idea that a new iteration might be a good time to sit back and consider where the project might ultimately go.

I suspect (without any actual knowledge) that the board was borne from 'I think I can do this, let me see if I can', which became 'My buddy down the road wants one', to 'There are a bunch of people on the Internet who want one', to 'OMG - this project has taken on a life of its own!'.  This might not be unlike the shack at the lake you built one weekend that you later decide needs an actual bathroom instead of an outhouse... then a second bedroom... and before you know it, you have a house built from multiple additions.  At about the fourth phase of cabin construction, you start to think about how you could have built the original shack to make your life easier when the ones that became necessary later were added.  I said 'suspect', because I've had my own projects that morph into something bigger than I expected, so I'm inferring... I'm also not suggesting that the current iteration is scabbed together - but I think you know where I'm coming from.

My guess is that the HL wasn't originally intended or expected to be what it has become.  At some point, probably when a new revision is considered, it might be a good idea to sit back and contemplate with a clean sheet of paper... If you were going to build the ultimate hacker/tinkerer SDR, what would it include?  Draw it all, even the stuff that might seem over the top.  Then go back to the core HL(3) which is your immediate project and consider how it should be revised to best allow for the future additions.  In the example of the cabin, framing in doors for the future rooms would have allowed you to just peel back the drywall and siding and carry on, without having to reconstruct the wall for the header, reroute the electrical up and over, etc...

I'm not saying this is necessary, but a proactive review never hurts.

73,

Brock
VA7AV

[John Williams]

You are actually talking about HL1. HL2 has a more pure soul. You underestimate the design and the designer. The core board’s full capabilities have yet to be explored. Partly because the core with the filter meets so many needs. With all open source, take the core in any direction that meets YOUR needs. Just an add-on board away.

My two cents.

John

W9JSW 

Sent from my iPhone

To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/ada8b443-c6e2-4359-8657-5714dc0ba838n%40googlegroups.com.

[Brock Nanson]

Hi John,

I actually tried to keep from insinuating anything negative about HL2 - exactly the opposite in fact.  I like it and have no reason to doubt it or the designer.   But I've learned the hard way that if you're going to make changes, think as far out as you consider reasonable... then go further.  I've saved myself time and expense in both personal and professional projects by planning for that extra contingency that seemed far-fetched at the outset.  It feels wonderful to be able to say in a meeting, "well, actually we CAN do that now, because we did this earlier, even though it didn't seem necessary at the time..." ;-)

I'm probably talking more about the physical form factor than the electronics or programming.  I'd like to make it easier for people to connect a variety of add-ons, improve heat dissipation, adapt an enclosure for different environments/purposes, etc.

Brock

[ron.ni...@gmail.com]

Brock writes:

> I'm not saying this is necessary, but a proactive review never hurts.

If you go back in this mailing list, you will find that "proactive reviews" already happened (more than once maybe?)

several years ago.  The HL2 is a great finished building block for multiple uses, not just a niche field day contesting box. 

The SMA connector is great for me, because one of my uses for an HL2 is as a high quality programmable signal generator.  SMA is far better than SO-239 for that use.  But I can add a SO-239 pigtail when I want to transmit out the big antenna.  Flexibility.

73,

Ron

n6ywu

[Steve Haynal]

Hi All,

I didn't mean to start rumors about a Hermes-Lite 3. Except for some possible gateware enhancements and bug fixes as well as the teensy-based keyer, I'm done with Hermes-Lite and related development. Others are welcome to organize here or fork and develop a Hermes-Lite 3. Everything is open source and hardware.

There is a lot of past development discussion in this group. The group began 7 years ago on April 19, 2014. We began with the original Hermes-Lite, which was very much a "try something quickly and see if it works" type of project. You can still see many of the original Hermes-Lite builds in the gallery:

https://github.com/softerhardware/Hermes-Lite/wiki/Gallery-of-Hermes-Lite-Builds

The Hermes-Lite 2.0 took many learnings from that. In fact, it took several years of part time work by several developers plus input from this group to develop the current Hermes-Lite 2.0. I've tried to capture contributions by other developers on the schematic:

http://www.hermeslite.com/

73,

Steve

kf7o

[Matthew]

Brock,

With a lot of what you describe, I see these as companion cards. There are already quite a few. As you infer, the options are endless. Furthermore, because the HL2 has been designed to be future proof and have many different options (study the schematics and you will see how forward thinking Steve and the team were - e.g. look at the PA, look at the clocks).

You perhaps may not be aware that there is a 100 W PA that closely couples with the HL2? This stacks above the HL2 and in *theory* could fit into the existing case. I am interested in coherent rx, so I have my HL2 setup moving into a small PC case. 48 V power supply modified into an ITX power supply case, x2 stacked HL2s and the 100 W PA on top. This is ongoing work for me.

I have never operated on 6m, so I set myself the challenge of designing a transverter the size of (and to replace) the N2ADR board. I'm nearly there with the PCB layout. This will use the HL2 clock output (again, good forward planning/open options by Steve et al.) to feed a standard diode ring mixer style xverter. I've design it so that one could populate for any band from 6m to 70cms. Again, ongoing work. See attached picture. Someone could of course put all the necessary filters on one PCB and switch between them, but I don't think it would fit on the N2ADR board footprint.

I hope more people come up with interesting companion cards. It is great seeing how people have taken the HL2 core and adapted it for their needs without needing HL2 PCB or gateware changes.

73 Matthew M5EVT.

[ron.ni...@gmail.com]

Hi Steve,

As long as this discussion has pivoted to potential planning ahead, how long do you think the current BOM will stay stable, before critical components become deprecated or no longer readily available?  (As has happened to previous OpenHPSDR designs).  When, out on the horizon, might new and affordable programmable logic, analog-digital, and/or other components, with significantly better performance (more gates, more MACs, more memory, more sample bits, higher sample rates, lower phase noise, better linearity, etc.), become available for something like an HL(n++)?

73,

Ron

n6ywu

[Pierre Martel]

Btw, after reading a bit more of this thread I started to dream about what would be the best hermes-lite for me.

Dual tuner, would be a good thing, with possibility to run both receivers for constellation reception. 

one TX port. 2 receive port. one of the receive ports could be switched on the TX port to do as we do now by a small reverse T/R switch. that can be added on demand.

Maybe upgrade the fpga. The radioberry already uses the cyclone 10 to run on a RPI. 

I wonder if this could be a foundation to a new era of great development. 

Not that the development has not been great already, but it could even be much more. ;-)

It would be fun to have a single port with all the IO that have not been used already or those that could be reused again beside the RF section available directly on one side of the board and the filter section already on the PCB.  The RF I would keep apart So that development could be done either on the full IO OR on the RF side without mixing them. Also the receiving filters could be put inline or offline for both receivers separately. 

Pierre

VE2PF

To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/f8d08e23-60c5-4353-b703-b03a9cda4712n%40googlegroups.com.

[Brock Nanson]

Steve et al,

I'm definitely sad to read that you're done with further development, but given the work you've put in to date, I fully understand and can only thank you for everything you've done. I'm new to this, but frankly can't imagine how many hours you must have invested!

Some of the comments back on my posts suggest to me that people are thinking I'm talking about the software and electronic design of the board.  I'm not, at least not primarily.  I have nowhere near enough knowledge or experience to do that!  When I suggest that some blue-sky dreaming about what might be added/connected as side projects would be beneficial, I'm thinking more about how to physically simplify electrical connection points and mounting in a chassis to benefit the development of those side projects... and the all-thumbs people like me who dare to attempt assembling them.

Some examples...

1. HL2 has holes in the board for mounting with standoffs.  N2ADR board does not.  So mounting in a different case, without the side rail support, means sorting out edge mounting for the filter board, ensuring heights match for the 'Lego' connector between the two boards... I'd rather use screws through both boards to make sure nothing is strained or allowed to float around.
2. If there were SMB connectors on the boards, instead of (or in addition to) the SMAs, a pigtail could be snapped in and run to chassis mount BNC, SO-239, SMA... whatever was desired, wherever makes sense for the builder.  Or an SMB jumper could connect RF to the next side project device.  Yes, an SMA pigtail could be used from the current connections, but it messes with the clearance to the other stuff that might need to be accessible/visible from the outside...and demands more internal space.
3. Heat sink for the finals that doesn't require the rail - sure, we could MacGyver something else to suit ourselves, but it's still an example of where a one-size-fits-all solution would simplify things.  If there's a desire for improvement, this should make the list.

So, yeah... I'm definitely not equipped to critique any of the technical stuff... don't allow any of my comments to suggest otherwise.  I'm looking more at potential incremental improvements to the practical build side of things.  I'm probably not qualified to do that either, but it's less of a stretch... ;-)

Brock

[Simon]

Hi all

I need to say this even though it may get me some abuse..sorry in advance.

Hl2 IS NOT for people who want a fully up and running plug  and play sdr..

Its for enquiring minds types WILLING to experiment and develop..

Its all open source..plenty of like minded people out there willing to share info..

The developer/s ( Hardware/ software quisk / sdr etc) have put in I would say divorce amount of time in it so far.( I have no idea how the poor man has a life outside HL2!) .leave him alone, work it out, most if not all issues have been covered on here. You wonder why the will not be HL3??? ( stand to be corrected here???) 

You want it painted blue, paint it blue.. 

you want it in a bespoke case with 1kw amp, then make it.. work out how to.. once you do post info for others..

You want a full 7 inch screen on front panel then buy something  else ( ftdx101? ) . Etc..

HL2 is a brilliant bit of sdr kit for the monies..( it is as good as my elad duo in the real world, though it should not be.)

NOW if the developer was running a ltd/ co company earning £££/$$ then yes ask for this/ that..but  this is not the case..

Learn from it..

BUY them now, as a certain chip is I suspect discontinued and they will not be around for ever..

Sorry for rant..been a hard day..very hard day..lost my best mate but he would agree with everything I said..he had one..

Regards Simon g0zen ( waiting for abuse/comments)

THIS WAS NOT aimed to anyone in particular!!  Just  saying what i think as at the moment I do not care ( see just lost best mate/ best man at wedding) and stating developer ( YOU ALL KNOW who  I mean.) probably needs a well earned break..I WOULD..

Best regards..stay well..happy sdr’ing

On 22 Apr 2021, at 21:07, Brock Nanson <br...@nanson.net> wrote:

Steve et al,

To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/e3aa9a5a-d4e4-45e9-8f2f-4e91b937d701n%40googlegroups.com.

[Pierre Martel]

Everyone is entitled to his opinion Simon! 

Sorry for your loss, Sometimes what we think is a simple wish and just saying what we hope could be done without any pressure on anyone can be sense as being a request. 

I think everyone can state their dreams. 

No one is obligated to anything. Maybe it can trigger a new passion in someone new. Maybe it can give inspiration to the original designer. cross my fingers. 

But in no way it should be taken as an obligation and if no one take the bait, I wont be sad ;-) 

I really like the HL2. I think everyone that helped to bring that marvelous radio up to where it is now deserves a lot of respect. 

And don't worry

No one will open the flame thrower so you don't need to wear a flame suit!

Pierre

VE2PF

To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/817A0CAE-39E9-4708-B994-A111F9329231%40gmail.com.

[Brock Nanson]

Heath,

That is a great video.  I might have to see how well my AutoCAD experience transfers to the free SketchUp package.  Snooping around Steve's GitHub files and related pages, there's likely everything one would need to play around with alternative enclosure configurations.  I started looking at the HL2 schematic PDF pages and comparing them to my board last night, but that's not something to try when you should really be heading for bed.... I went cross-eyed pretty fast - LOL

I don't particularly want to open up the heat discussion again, but I would like to understand how heat is dissipated from the FETs themselves.  Remember, my knowledge of this stuff is minimal... Looking at the technical specs for them, I infer that heat is intended to be extracted via the center 1/3 of the bottom, where the metal tab runs across the FET body and narrows to become the lead.  The metal block we're using is on the opposite side of the board, so I'm guessing heat must pass through the board substrate to reach the metal block.  Is that right?  If that's correct, then would there be any advantage to using some sort of extruded aluminum piece on the top of the FETs, carefully secured with a screw through the hole between them?  In addition to, or instead of the metal block?  I'm wondering whether the FET case would transfer heat better than the board substrate.

Brock

[ron.ni...@gmail.com]

Hi Brock,

Judging from the fact that the HL2 ran almost as cool before the heat shim block was included, a lot of the heat is likely dissipated through the plated through-hole vias and bottom of the PCB to the side of the metal case.

There's tons more info on heat dissipation calculations and measurements on this mailing list many months ago.

73,

Ron

n6ywu

[Steve Haynal]

Hi Group,

Just a few comments on this thread.

I encourage people to dig through this list and the wiki to try and find answers to questions. Many ideas which come up have already been discussed in the 7+ years of this list.

* Heat dissipation for the LDMOS devices are to the inner PCB layers and through to a bottom heat spreader (the rail in most cases). Details of this information can be found in the datasheet and application notes for the LDMOS devices. You can find several discussion covering this during the design phase of the HL2. The HL2 board already has a large bare section under the PA for heatspreader experiments. The heat shim is one such experiment, but others are possible.

* Heat is produced in a PA when the power supply voltage is higher than the output signal voltage at that instant and the transistor has to "burn" this excess voltage at the required current. One way to learn about this it to understand how EER techniques can reduce heat. See this thread and the lecture notes link:

https://groups.google.com/g/hermes-lite/c/hsDmvY6zxyE/m/XAOV45t5CgAJ

https://github.com/softerhardware/Hermes-Lite2/wiki/Enabling-EER-PWM-Output

By the way, adding EER to the HL2 is a very interesting and worthwhile project for someone who wants to reduce heat. It is very much in line with the software side of software defined radio. It is one of the novelties of the ucx:

https://groups.io/g/ucx

* For cost reasons, the HL2 has only one ADC and DAC. There are experiments to link to HL2 for phase coherent RX and TX. Search the group for "diversity" "coherent" "synchronized." See this thread and wiki page. This is an area where I still plan some gateware work.

https://groups.google.com/g/hermes-lite/c/FYS15G-tWhs/m/T7VLmj1RAQAJ

https://github.com/softerhardware/Hermes-Lite2/wiki/External-Clocks

* There are already several surface mount 50Ohm uFL connections sprinkled across the HL2 board. See this options section and the schematic:

https://github.com/softerhardware/Hermes-Lite2/wiki/Options#rf-and-clock-connectors

https://github.com/softerhardware/Hermes-Lite2/raw/master/hardware/hl/hermeslite.pdf

* I have considered the Cyclone 10. In fact, I either suggested it to Johann or he saw the early discussion on this list. The closest pin compatible part has 3 less IO than the Cyclone IV we currently use, so would be hard to use. Plus, it provides only ~5% more gate resources. I asked Makerfabs if it made sense to switch to a smaller BGA FPGA and shared with them the difference in part cost. They did not think it was worth the effort and risk. Also, I think Makerfabs has a good deal on the current FPGAs. You can find much discussion about FPGAs on this list. Search for Lattice, Xilinx, ECP5, qmtech.

* The N2ADR board can be modified if there is something not liked, or even an entirely new board created as the HL2 and filter is a modular design. You can find all the design files for the currently manufactured N2ADR board here:

https://github.com/softerhardware/Hermes-Lite2/tree/master/hardware/companions/n2adr

The open source tool kicad is used for all schematic and PCB design, so even the tools are available to make modifications:

https://kicad.org/

It is easier and easier to have custom PCBs made, even with some assembly:

https://oshpark.com/

https://www.makerfabs.com/pcb-prototyping.html

https://jlcpcb.com/

* Some parts on the HL2 BOM will probably become obsolete in the next few years. The heart of the project, the commodity AD9866, is already not recommended for new design. I think it was introduced in 2013, 18 years ago! I do like to look for new alternatives and have yet to find another part with the right cost, performance and LNA integration as the AD9866. Newer parts are either too specialized, too high end or require too much surrounding support. The PA LDMOS devices are also nearing end of life. There are better alternatives for the FPGA, but the benefits of changing that in the current design are not worth the effort. For example, Xilinx just introduced their Kria modules, and the upcoming commodity module might make a nice platform for a next generation SDR:

 https://www.xilinx.com/products/som/kria.html

This reply has become very long, but in summary, I would say that the HL2 design solidified several years ago. People are encouraged to experiment with their HL2, but it is unlikely that the core HL2 design will change. Makerfabs will continue making and selling the HL2 as long as there are buyers. (We've already exceeded expectation here by >3X.) If some people do develop a companion card or even upgrade the HL2 in a widely desirable way that adheres to the collaboration guidelines:

https://github.com/softerhardware/Hermes-Lite2/wiki/Contributor-Guidelines

I may be willing to help have Makerfabs produce it for a wider audience, provided that others are doing the bulk of the R&D work.

73,

Steve

kf7o

[Josh Logan]

This has been an interesting thread to read.  I know I'm 2+ months late, but still wanted to share my experiences.  I have been spending a lot more time operating and a lot less time reading about radios.

The HL1 was quite the hobbled together system, and I never even got my 2 band PA for that built before HL2 was announced and I put those parts in a box.  But HL1 really did excite people to the idea of multiple live receivers at the same time and a group of software developers ready to jump in and create unique and usable software that supported the new feature set.

HL2 really did have some initial teething pains but  those were wrapped up quickly and then the creation of the N2ADR filter board was complete and the platform stabilized.  From build 5, or so, it really did stabilize into a building block that has provided me years of enjoyment.  

This thread was initially around temperature and I have many times been transmitting for too long trying to get a loop to tune, or being connected to the wrong antenna.  The TX heat protection has always just paused transmitting until the heat dropped again and then it would pick right back up.  I use the 55mm case, no extra shim.  Might there be a time/place to add a fan, maybe, but experience has shown it's also quite tolerant to operator error.

Forward looking....  This HL2 hears most FT8 stations better than my Kenwood TS-2000, and I never get messages saying it splatters.  I have multiple bands monitored at once so I know where to spend my time operating.  I look forward to seeing more progress on the 100W PA, and transverter boards.  But most of all I appreciate the whole community that has built a radio, and then a system, that has gotten me operating more and more.  I have gone from some UHF/VHF contacts and a handful of voice contacts to over 2000 FT8 contacts, Worked All States on 3 bands, and 91 confirmed DXCC entities.  All from a low hanging 40/20/15 fan dipole.  Not all with 5 watts, and not all with the HL2, but all because of this little radio encouraging me to make contacts.

73, Josh

KD7HGL

To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/264a2201-8b8a-4161-9b2b-caf5c7300069n%40googlegroups.com.

[didier....@gmail.com]

Hello guys,

According the final temperature , i simply use the initial mod with a hole of the casing and a screw/bolt to maintain firmly the bottom of the PCB against the casing  In addition i just add a 5vdc little fan on the top of both FET . The temperature is hardly above 45c at full power even when using PSK31 or FT8 modes. The main benefit is after a long duration QSO likely PSK31 mode , the FET temperature is falling down very quickly to nominal.

i  am using 2 hermes lite (one is the initial one) and both are equipped with a fan and i do not care about the FET temperature

As mentionned by many guys , the Hermes lite is open source and mostly a wonderful  platform which can be used  for your own mods , integration, or enhancement.

73s F5NPV

[Steve Haynal]

Hi Josh and Didier,

Thanks for the kind words. One of my HL2s is in my sister's garage. I was using it even for FT8 transmit on the first day of our recent heat wave and still did not see the temp go above 50C. This one does have the heat shim installed. I took that system down for computer upgrades, so don't know how it did in the height of the heat wave. I'm setting up a computer with a better processor so that I can decode FT8 at depth 3 on all 10 HF bands with SparkSDR.

73,

Steve

kf7o