HL2 Interrface to Filters and Other Boards

[James Ahlstrom]

Hello Group,

I would like to be sure that the HL2 design accommodates boards other than just the 5x10 cm filter.  I encourage other board designers to contribute to this topic.  I have paper napkin designs for boards with multiple Rx/Tx filters, VNA boards and transverter boards.  By throwing them all into the design now, it will help to make sure the HL2 interface is as robust as possible.  The logic connection at CN7 etc. is addressed in a different thread here.  Let's look at the RF connectors.

There are three RF connectors at the back of HL2.  RF2 is the 5 watt output and Rx input and is connected to the T/R relay.  It can also be used as an Rx-only input under firmware control.  RF3 is an optional Rx input that is in parallel with RF2 Rx, although since B82 is not installed it is currently unused.  RF1 is the low power Tx output.  The op amp driver is switched between RF1 and the finals under firmware control.

I believe there is a problem with RF3 caused by its connection to the T/R switch.  Let's look at some boards.

HL2 Only

------------

Consider the no board case.  We just connect the HL2 to two RF connectors on the back panel with short coaxial cables.  We can't use RF3 because the extra cable capacitance loads RF2 on Rx.  If we want a separate Rx input, we would omit the T/R relay and wire directly to the relay pads using the extra grounds Steve provided for the purpose.  RF3 remains unused and we now have a dedicated RF input from the relay pads.  OK so far.

VNA

------

Let's put the HL2 into a 10x12x3 cm Hammond 1455L1201 box, and design a 10x2 cm board with an internal bridge for VNA.  We need need 4 or 5 connectors; namely the VNA generator and detector for transmission measurements, the VNA unknown port for reflection measurements, and one (with T/R) or two (separate Tx and Rx) for normal HL2 operation.  We can wire RF2 directly to the back panel and use RF3 and RF1 for the bridge.  This requires a PE4259 switch directly at RF3 to avoid loading RF2 when the VNA is not in use.  It is a problem that the user must disconnect RF2 when using the VNA to avoid loading RF3.

Let's go back to a 10x15 cm box and use an external bridge.  We can use RF2 for the detector and RF3 for the generator.  But if we are using a filter board, the filters are in the way.  And if anything goes wrong, the 5 watts will burn out the bridge.  We could use RF3 and a PE4259 as above, but the filter board loads the bridge.

Filter Boards

-----------------

The 5x10 cm filter board is in good shape provided the 3 MHz Rx HPF is on the HL2 board, but I believe that any Rx high pass filters should be on the filter board.  I could go back to switching out the HPF on Tx, but a better design is to move the T/R relay to the filter board.  Other filter boards with multiple Rx filters will need to have their own T/R switch and a separate Tx and Rx path to the HL2.

Transverters

-----------------

These require RF1 and a dedicated Rx path.  Using RF2 for Rx is dangerous because if something goes wrong, the 5 watts will burn out the transverter.  It would be better to use RF3.

External Amplifier

-----------------------

If we omit the filter board and use HL2 as a driver for an amplifier chain, we will need a separate Tx and Rx connector, and the T/R relay will not be used.

None of the above boards are benefited by having the T/R switch on HL2 except that it saves a little space on the 5x10 cm filter.  I would like to try designing a 5x10 cm filter with the T/R relay on the filter and still including the 160 meter band and a single Rx filter.  So I would like to remove the HL2 T/R relay.  The relay driver remains connected to CN12.  The extra space can be used for UFL connectors, although I have to admit I don't really know what those are.  RF2 is 5 watt output and RF3 is Rx input.  An even better design would be to add a fourth connector for alternate Rx input switched with RX3 by a PE4259.  It is easy to find space for an extra PE4259 but not easy to find a spare IO pin and routing to switch it.  So I would just connect the switching to U10 and U11.  That means RF3 is the Rx input when RF2 is in use, and the alternate Rx is used when RF3 is in use, a very simple rule.  This is very handy for the VNA and transverter, and I think it is what we were originally aiming for with RF3.

What do other board designers think?

Jim

N2ADR

[Dani EA4GPZ]

El 26/09/17 a las 20:06, James Ahlstrom escribió:

> External Amplifier
> -----------------------
> If we omit the filter board and use HL2 as a driver for an amplifier
> chain, we will need a separate Tx and Rx connector, and the T/R relay
> will not be used.

Hi Jim,

Just to point out that I currently use the HL2 beta2 with a Hardrock
50HF amplifier using RF2 for both TX and RX. This amplifier (and others
design to hook up easily to something like an FT817) has a single TX/RX
connection.

So in fact I do use the T/R relay in the HL2 board. If I had a filter
board for the HL2 with a T/R relay, I could connect the HL2 to the
amplifier through the filter board and use the T/R relay in the filter
board. However, in my case a filter board is not needed, as the
amplifier output is filtered.

It would be nicer to have different connections in the amplifier for TX
and RX, but that would make it more difficult to use it with the FT817
(which I also use with this amp).

There is something in the current design that I find a bit strange. I
have B82 installed and sometimes use RF3. However, in TX the T/R relay
shorts RX to ground even if using RF3. I don't know what was the
rationale for designing it in this manner. Perhaps it would be better to
connect B82 in such a way that when the T/R relay is active, RF3 is
still connected to the RX input of the AD9866.

73,

Dani.

[Graeme Jury]

Hello Jim and Group,

Thank you Jim for a timely review of the HL outputs and switching and yes I agree that for me there are some issues which could be directly addressed or else I will need to do a workaround on connected boards. To my way of thinking an exciter/transceiver should supply RF out and Rx in with separate inputs plus a set of control outputs and monitoring inputs. I will also put aside the control and monitoring issues and concentrate on the RF side of things.

The Hermes-Lite board meets my general requirement if I omit the changeover relay and use RF1 as my low power Tx output, strapping RF2 to my Hi power Tx output and using RF3 as my receive input,

HL2 only

I would run the board into a filter board with a changeover relay after the filter and mounted on the external filter board. The reasoning is that if I provide any 1.7 MHz HPF filtering (which I need) I would not want to run the Tx through it as hopefully I can do it with small smd components. I believe that the 30 MHz LPF belongs in the HL board on the receive leg for nyquist purposes and allows the HL to run on Rx without any external filter board.

VNA

In an ideal world the vna out and in would be switched with a pair of PE4259's so that the HL board had 5 RF connectors with Lo power Tx and Rx in being on one pair and auxiliary Lo power Tx with auxiliary Rx in via the PE switches being the other pair. Hi power Tx is on the 5th connector and not switched. To be realistic it is probably too much real estate for HL board to fit that many RF connectors plus the PE4259's so this could easily be moved to the filter board if using 100 x 100 mm but probably impossible on 100 x 50. The VNA would connect to the auxiliary RF connectors.

Filter Boards

The requirement for the filter board has been established by the preceding HL and VNA discussion. I expect a few filter designs to emerge based on the intended use of HL ranging from a whisper band hopping station to a full blown KW output contest grade transceiver. In the case of a top end rig HL will probably drive an intermediate amp in the 10 to 20 watt class and the filter will reside after that and just the Hi power Tx out from RF2 and Rx in from RF3 would be used. This is where separate switching of the vna signals onboard on HL would be so valuable. In my case if the solid state switching works out (still looking promising) then the high speed T/R switching on the filter board would offer an additional advantage.

Transverters

Yes RF1 would need to be used here unless the transverter was of unusual design and needed 5 watts on Tx leg. Again the removal of the HL changeover relay would be required and Rx on RF3 connector. It would also be possible to use the auxiliary RF connectors if they happened.

External Amplifier

This is covered in the previous discussion and to reiterate the amplifier would be driven from the RF2 connector which would be in conjunction with removing the changeover relay be strapped to B98 on the output transformer. Receive would again be on RF3.

In all cases the changeover relay is a disadvantage for my personal use of HL and would be removed if fitted but I can see that for somebody running HL barefoot with a shared Tx and Rx filter it would be a very compact and clean design. The relay driver should remain connected to CN12 as that could affect a number of possible output boards. I would be interested in hearing the opinion of those who would feel that having the T/R relay on the HL board is the best approach and if their concerns could be met by an output relay board which possibly holds an over voltage protector or other devices if they don't want it on a filter board.

I have in the past connected between boards with .1" header and ribbon cable with the RF wire having an earth each side and it seemed to work just fine. The run was only about 40 mm though.

73, Graeme zl2apv

[Graeme Jury]

Hello Dani,

You have answered my question although if you had used Jim's proposed filter with a T/R relay you would have the same result with the advantage of a better filtered Tx drive signal.

73, Graeme zl2apv

[Steve Haynal]

Hi Jim and Group,

Populating the TR relay on the main HL2 board has always been an option in my mind. With the last build of Hermes-Lite beta3, this relay was not stuffed by the assembly house. Like the power connector, RF connectors, and other front and back connectors, the TR relay was not included in the assembly BOM so that a user could customize. I've always known that some would not like a slow relay, or even any relay, at this point in the RF chain. So, if you can fit a TR relay, RX HPF and 160M-10M TX LPFs on your 5x10cm companion card, then go for it, it sounds excellent.

I do not want to remove the TR relay option on the HL2 board though. Users like Dani and myself (I like to sometimes use an inline filter or filter bank with no TR) find use for it. As engineered, there are three configurations for RF connection on the HL2:

* K2 TR relay installed on HL2

Only when the PA is enabled does the onboard TR toggle. When in low power mode, the TR always stays in the default RX mode. This is to support full duplex applications with low power TX out on on RF1 and RX in on RF2. This is why there are two TR relay drivers, CN11 and CN12. CN11 always toggles on TR. CN12 only toggles on TR when the PA is enabled. CN11 is used as external TR when running in this full duplex mode. CN12 may never be used externally, but since there was space, I made it available. This may change slightly in the future.

For VNA operation, it is expected to connect a temporary external bridge. Only basic transmission and reflection measurements as possible by the original Hermes are intended to be possible with the HL2. (Only S11 and S12, and different external attachments may be required for S11 and S12.) Only low power is necessary and available for VNA operation.

Optional RX input RF3 doesn't have an intended use in this configuration and is expected to be left off. RF1 and RF2 are spaced far enough to support large coax and PL259 connectors. RF3 will just get in the way of that. I'm surprised that Dani reports this is shorted to ground sometimes. I will have to investigate.

In this configuration, it is possible for someone to connect RX in to RF2 (RF2 is RX in for low power full duplex) but not disable the PA and transmit into whatever is connected at 5W. To guard against this, most software does not enable the PA by default. Any VNA software should not enable it. I actually prefer gadgets to not have too many safeguards as they sometimes get in the way of experimenting. I am willing to convert CN8, CN9 or CN10 (run out to the edge in the next version) to a PA inhibit. Any external board VNA or transverter boad that does not want the PA to accidentally turn on can ground one of these and the PA will be disabled in firmware. The PA is already disabled in VNA mode or low power more.

* K2 not installed on HL2

This is the configuration intended for those who desire some other form of TR switching, such as pin diode for QSK CW or more flexibility as in your filter design. Users don't install K2, but must install a small wire jumper between pin 8 and 9 of the K2 footprint. They must also install B82. I will try to convert B82 to a 0805 or larger footprint, maybe jumper, to make this easier. 

In this configuration, RF1 is low power TX out, RF2 is high power TX out, RF3 is dedicated RX in. VNA attachments are expected to connect to RF1 and RF3 and provide only functionality as previously described. CN11 is again for external TR. 

* K2 not installed on HL2, HL2 supports receive only

In case someone wants to build a batch of HL2s just for RX, then the associated TX parts (preamp, PA, relays, etc.) can be left off. B82 is stuffed and RF3 is dedicated RX in.

   

73,

Steve

KF7O

[James Ahlstrom]

Hello Steve,

Thank you for clarifying the design.

On Wednesday, September 27, 2017 at 12:02:44 AM UTC-4, Steve Haynal wrote:

* K2 not installed on HL2

This is the configuration intended for those who desire some other form of TR switching, such as pin diode for QSK CW or more flexibility as in your filter design. Users don't install K2, but must install a small wire jumper between pin 8 and 9 of the K2 footprint. They must also install B82. I will try to convert B82 to a 0805 or larger footprint, maybe jumper, to make this easier. 

I am troubled by the need to install B82 even if it is 0805.  For the design you describe it is not necessary to wire RF3 to the T/R relay.  RF3 can just be connected to a through hole pad near the relay.  Then the user wires the relay Tx to RF2 using the pads as you describe, and wires the relay Rx to the RF3 pad.  We can get a bit more space by eliminating B82 because the DC block is provided by B81.

Jim

N2ADR

[Steve Haynal]

Hi Jim,

Okay. I will try to try to convert to a wire jumper here. I opened an issue for this last night: 
https://github.com/softerhardware/Hermes-Lite2/issues/66

73,

Steve

KF7O

[Dani EA4GPZ]

El 27/09/17 a las 06:02, Steve Haynal escribió:

> Optional RX input RF3 doesn't have an intended use in this configuration
> and is expected to be left off. RF1 and RF2 are spaced far enough to
> support large coax and PL259 connectors. RF3 will just get in the way of
> that. I'm surprised that Dani reports this is shorted to ground
> sometimes. I will have to investigate.

Sorry. I was commenting from memory. Reading the schematic again, it
seems that when the T/R relay is on, RF3 should still be usable as an RX
input. I've read the current schematic, but I think this part of the
schematic hasn't changed since beta2.

By having the T/R relay on continuously (I think there is not support
for this in the current software/firmware), one could use RF3 as RX
input and RF2 as 5W output, even in full-duplex.

73,

Dani.

[Takashi K]

Hi Steve and all,

I think there is restriction on use case, that when using an external amplifier (such as HARDROCK50) and the post RF pick-up with RF2 connection, HL2 does not have Tx monitor input for pure-signal.

73, Taka  ji1udd

[Daniel Estévez]

El 27/09/17 a las 23:23, Takashi K escribió:

Hi,

I think that you could use RF3, though I haven't checked it. I thought
that there was enough RF coupling to monitor your TX in full-duplex, but
now I'm not sure. It seems that if you're using a external amp such as
the Hardrock, most of the coupling you get is from the onboard 5W PA, so
you don't know what your amp is doing (in terms of IMD and such). I
still have to check this.

73,

Dani.

[Takashi K]

Hi Dani,

> most of the coupling you get is from the onboard 5W PA, so
> you don't know what your amp is doing (in terms of IMD and such).

I guess so,too. Therefore I guess Pure-Signal doesn't work perfectly.

This is one of use cases, not my use case. I hope you check this.

73, Taka  ji1udd

[Steve Haynal]

Hi Dani and Taka,

Good point. If we keep TX enabled, we can have full duplex with PA and RX on RF3. I've added this to the github issues for future implementation. Also, we can recommend that full duplex always use RF3 for RX even with low pow full duplex. If no load is attached to RF2, then RF3 should always work well for RX no matter what state K2 is in. This also avoids the danger of accidentally driving the receive side of a transverter with 5W...

Not much has been done regarding pure signal and an external amp. It may be possible and necessary to tap some of the external amp output, attenuate and feedback to the HL2 via RF3.

73,

Steve

KF7O