HL2 Test PA

[Steve Haynal]

Hello Group,

I am finishing up a test PA for HL2. You can find the KiCad files as well as schematic (pa.pdf), bom (pabom.pdf), layout picture (pa.png) and adapter board picture (adapter.png) on the HL2 github fork and attached to this post. The purpose of this board is to test the layout for HL2 as well as some new PA build options (push-pull, non TO-220 cases with adapter boards). I want to gain confidence in the design and layout before committing it to HL2.  Claudio has kindly offered to test and refine the HL2 PA. I need to order the PCBs to be sent to him very soon so that he receives them when he has some free time, so please provide any feedback in the next few days. Some points about this design:

• It is meant to be in the back right corner of the HL2. The TO-220 cases bolt to the side of the target enclosure for thermal dissipation. The adapter boards for other transistors may also bolt to the side and connect to the HL2 via very short copper wires or braids. I considered the mirror image (back left corner) and may switch to that, but given the pinout of the AD9866, this back right corner orientation is most likely.
• The RF in comes from the front.
• The RF out goes to the left back corner where the filters will be. It is a tight fit give the size of the relays, but the PA layout is narrow enough.
• This can be built as class A with a single transistor, or push-pull.
• John's tested class A PA (minus the choke on the power) can be built with this "universal" design. I left the choke off as the RS-HFIQ and other designs with a similar PA do not use one.
• This is very much a test design and I don't recommend that anyone manufacture PCBs to build. There is still much refinement to do. Any other interested and capable testers can contact me or Claudio for a spare board. 

73,

Steve

KF7O

[John Williams]

Steve, et.al.

This looks like a very good plan. With Claudio's help, I am sure you will see sure success. Design looks good to me. We are still moving forward on the filter designs on the 1.1 board. I just sent one to Stew so we can get his measurements on the board as well. With your move to the PA on the main board, I will probably not plan to make any more versions of the PA, but can make them available along with the 1.42 frontend to anyone that wants to have a functional PA on the 1.2 design.

The RX board will need another revision. I made some logic errors in the switching that need to be corrected.

John

--
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.
For more options, visit https://groups.google.com/d/optout.

[Takashi K]

Hi Steve,

FYI,

If your PA target is 7.2V ( 2cell Li-ion ) operation and SOT89 package, I know another device, Toshiba RFM04U6P.

Mouser, Data Sheet, Application Note.

73 Takashi, JI1UDD

[in3otd]

Thanks Takashi,
I'll add that to my next Mouser order, together with the other RF devices to test on Steve's PA board.

73 de Claudio, IN3OTD / DK1CG

[James Ahlstrom]

Hello Steve,

I looked at this design for several days, and I don't think I have much to add.  It is a very good general design, and can be made into several versions by interchanging the 0805 parts, as you say.

The possible problems of heat dissipation and the signal source integrity for the push-pull version are hard to address without the whole HL2 being completed.

Jim

N2ADR

[Steve Haynal]

Hi Jim,

Thank you for looking over this design and providing feedback. I have updated github with the (hopefully) final design and plan to submit to elecrow in the next hour or two so that they can being first thing Monday morning. Here are the Gerber files for the adapter and testpa. I use Gerbv to view these files on either Windows or Linux.  

Since the last schematic and png post, Claudio found a bug with pins 4 and 5 of T1 swapped in the schematic versus the PCB footprint. It is fixed in the latest files. My biggest concerns are ground planes under the transformers and unsymmetric route lengths for the balanced feeds to the two transistors. I have removed ground planes and minimized signals on the two layers below the transformers to buy an extra 1mm+ of space, but need to run signals on the bottom to be compact, so there are signals on the bottom with a ground plane just above. I think this is less of an issue for transformers versus inductors and also for the HF RF frequencies we are working at. Also, given the placement constraints, it is hard to balance the differential signal lengths to the two RF transistors. Again, I'm hoping the difference in route lengths (roughly 2cm versus 4cm) will not matter for HF.

I'm sure once Claudio begins experimenting with these boards, he will find other issues and opportunities for improvement.

73,

Steve

KF7O