Awesome Idea the si4732 integration 🥇

[iu0azb]

Hello phdlee, I discovered your firmware here on github almost by chance. since I have a k5 dedicated to testing and a si4732 in the drawer, I'll try to put it in place of the original chip! I just have a doubt about the work around the chip.. Looking at the shematic around the actual BK1080, it would make sense bypass or adapt the Q7 circuit.. and replace C28?
Last, if you could share a picture of your PCP modified it could be helpful to have a clear idea :) Sorry to have used the issues form but it was the easiest way to contact you :)

[phdlee]

I worked on SI4732 MOD a few months ago. I didn't have a circuit diagram at the time, but I really like the circuit diagram you sent me. I checked whether the clock occurred in OSC1 to reduce the number of parts, but was not successful. So, three more parts were used. (X-TAL, CAP 2) If I had seen this circuit diagram earlier, I think I would have challenged myself a little more LOL In my opinion, the Q7 should be there. It acts as a PIN diode. There is a (FM RADIO ANT) RX PIN diode on the front of the Q7, so there is no big problem, but the radio designer probably took interference with the BK4819's RX IN into consideration. I installed a CAPACITOR on the Q7's BASE and connected it directly to SI4732. It didn't seem to make much of a difference, so I kept it back to its original state. If you are going to give up FM radio, it would be a good idea to replace c28, l9, and c29.

I investigated OSC1's clock and according to the BK1080's manual, it said that BK1080 does not need a clock, so I determined that OSC1 did not generate a clock. If this circuit diagram is correct, it would be a good idea to study OSC1 a little more.

DE KD8CEC

On Sat, Jan 20, 2024 at 2:28 AM iu0azb @.***> wrote:

Hello phdlee, I discovered your firmware here on github almost by chance. since I have a k5 dedicated to testing and a si4732 in the drawer, I'll try to put it in place of the original chip! I just have a doubt about the work around the chip.. Looking at the shematic around the actual BK1080, it would make sense bypass the Q7.. and replace C28? Sorry to have used the issues form but it was the easiest way to contact you :)

BK1080.JPG (view on web) https://github.com/phdlee/uvk5cec/assets/63300383/14ffe2e1-cd40-4ab6-b5aa-d12c29cbf73a

— Reply to this email directly, view it on GitHub https://github.com/phdlee/uvk5cec/issues/6, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE4DZPAAVOLSSWYJFYIFODDYPKUK3AVCNFSM6AAAAABCCJD6GCVHI2DSMVQWIX3LMV43ASLTON2WKOZSGA4TAOJXHAYDENA . You are receiving this because you are subscribed to this thread.Message ID: @.***>

[iu0azb]

After sent the message I decided to attempt a frankestein to have a look into the factibility of it... and after some tentative I won :) Here attached the working monster :)

About BK1080 pin 1 is getting 26mhz reference (200mV) .. too far from 32.768khz from doc it appear cleare that should not work.. but better to doublecheck :)

[phdlee]

Thanks for the interesting research. If 26Mhz is measured, it will be difficult to use. In the code that controls the bk1080, there are values that are entered into the internal area that are not disclosed to the general user, so unlike the bk1080 manual, it is not possible to know whether other frequencies can be input.

1. The code says that an external clock is used.
2. The manual says to use a 32.768Khz clock. However, 26Mhz is measured (bk4819 source clock) 3.bk1080 register addresses 0x16~0x32 are not disclosed, but specific values are put there in the source code. The source code has been reverse engineered, so even those who release the source code may not know what it means. This is the conclusion lol

Thanks for the measurements.

On Sat, Jan 20, 2024 at 6:42 AM iu0azb @.***> wrote:

After sent the message I decided to attempt a frankestein to have a look into the factibility of it... and after some tentative I won :) Here attached the working monster :)

About BK1080 pin 1 is getting 26mhz reference (200mV) .. bit far from 32.768 but not sure this reference change can be adjusted from settings.

— Reply to this email directly, view it on GitHub https://github.com/phdlee/uvk5cec/issues/6#issuecomment-1901153813, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE4DZPAFYDADYYZKVLVGO6TYPLSFRAVCNFSM6AAAAABCCJD6GCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMBRGE2TGOBRGM . You are receiving this because you commented.Message ID: @.***>

[iu0azb]

Yeap I'll try to investigate in codes too

For si4732 there is a pre scaler setting as detailed in pu2clr pages

https://github.com/pu2clr/SI4735/blob/master/extras/schematic/README.md I suppose that enabling external clock and using a prescaler set to 800 it could work

si4735.setRefClock(32500);
si4735.setRefClockPrescaler(800); rx.setup(RESET_PIN, 0, POWER_UP_AM,SI473X_ANALOG_AUDIO, XOSCEN_RCLK);

[phdlee]

I think you found the answer. I am so happy. It is specified on page 35 of the 4735 manual. I can try it out right away and, if possible, reduce 3 useless parts.

On Sat, Jan 20, 2024 at 3:54 PM iu0azb @.***> wrote:

Yeap I'll try to investigate in codes too

For si4732 there is a pre scaler setting as detailed in pu2clr pages

https://github.com/pu2clr/SI4735/blob/master/extras/schematic/README.md

— Reply to this email directly, view it on GitHub https://github.com/phdlee/uvk5cec/issues/6#issuecomment-1901806380, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE4DZPFNF4BPOXXNTLYE7OTYPNSZLAVCNFSM6AAAAABCCJD6GCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMBRHAYDMMZYGA . You are receiving this because you commented.Message ID: @.***>

[iu0azb]

Great ;)

[iu0azb]

Hi Seongho, today on a work break I've reopened your website, noticed that you updated the article. Thanks for mentioning me :) However, I noticed that you may have tried to use the prescaler without success. Do you confirm? Hopefully if Iyou want share the bin file I could do some tests to overcome the reset problem. No rush because I'm currently looking for the best solution to place the chip.

In the mean I'm trying to find the MCU section schematic.. eventually a dedicated reset pin can be setup by software..

[phdlee]

all right. I will send you a binary that allows you to use it as a prescaler.

On Tue, Jan 23, 2024 at 5:37 AM iu0azb @.***> wrote:

Hi Seongho, today on a work break I've reopened your website, noticed that you updated the article. Thanks for mentioning me :) However, I noticed that you may have tried to use the prescaler without success. Do you confirm? Hopefully if I could share the bin file I could do some tests to overcome the reset problem. No rush because I'm currently looking for the best solution to place the chip.

— Reply to this email directly, view it on GitHub https://github.com/phdlee/uvk5cec/issues/6#issuecomment-1904764880, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE4DZPF7RT3XCJH5DKDW3J3YP3EX7AVCNFSM6AAAAABCCJD6GCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMBUG43DIOBYGA . You are receiving this because you commented.Message ID: @.***>

[iu0azb]

I'm also trying to find the MCU schematic .. if an I/O pin is not used it can be used to reset the si. This should simplify the integration using common libraries

[jh4vaj]

The circuit diagram is on GitHub.

https://github.com/ludwich66/Quansheng_UV-K5_Wiki/tree/main/hardware/KiCad/Kicad_UVK5_reversing20230909

According to CPU.kicad_sch, it seems that PB15 and PC3 are unused. PB15 is pin-25, and since it's at the corner of the QFP, it might be easy to connect a wire.

I noticed that the MPU is mounted on the back of the PCB. Unfortunately, it seems difficult to route the wire around.

[iu0azb]

That's great.. At the moment I've installed the si4732 on the K5 PCB with RCLK connected to ex pin1 of bk1080.. Ready to test bin file once you share 👍 Reviewing si4732 doc, the use of prescaler should not be correlated to reset circuit .. but clearly I've a lot to study on it :)

[phdlee]

I wrote an email, but it wasn't all sent due to capacity issues, so I'll send it in several parts.

iu0azb, jh4vaj According to the datasheet, it looks like PC3 can output a clock from a timer. I can't be sure because I can't see the circuit diagram right now (it's somewhere other than my computer), but I'll check for recycling of the SWD port tomorrow. today I found out a few things

1. I worked on sending the firmware to IU0AZB. I had to be careful because I had to send firmware that didn't work in my environment. I looked at the code to make sure the si4732 library was working correctly and found a few unknown bugs. si4732 has two pins that receive clock input, and I took turns testing them. However, as you can see in the library source, that part is commented out.. All the DCLK pins I tested were unnecessary lol If the 12th BIT is 1, it is DCLK, and if it is 0, it is the REFC pin. It seems to have been commented out because it does not work while shifting 1 to the left 13 times. Anyway, I have checked all reference clock related parts, almost completed the firmware for 26Mhz clock, and completed all checks. I'll send it tomorrow, (checked firmware bin) and I'll send it to everyone who needs it. please send me your email address @.***

[image: image.png]

On Wed, Jan 24, 2024 at 11:54 PM iu0azb @.***> wrote:

That's great.. At the moment I've installed the si4732 on the K5 PCB with RCLK connected to ex pin1 of bk1080.. Ready to test bin file once you share 👍

— Reply to this email directly, view it on GitHub https://github.com/phdlee/uvk5cec/issues/6#issuecomment-1908292997, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE4DZPDQEJXFZFSGKWRM2FLYQEOCJAVCNFSM6AAAAABCCJD6GCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMBYGI4TEOJZG4 . You are receiving this because you commented.Message ID: @.***>

[phdlee]

2.I found out why it didn't work. It seems the clocks are not satisfactory. That's because the clock size is small and the bias is at 0 volt. In uv-k5, the clock comes through cap. It is the cap on the left of pin 1 of the bk1080. As it passes through the cap, the bias becomes 0volt. It seems that bk1080 can operate that way. However, when I experimented with si4732-a10, it was the same as a normal digital clock (1/2 of vcc).

The photo below shows the minimum operating conditions. [image: image.png]

It started to operate when the above conditions were met. [image: image.png]

On Wed, Jan 24, 2024 at 11:54 PM iu0azb @.***> wrote:

That's great.. At the moment I've installed the si4732 on the K5 PCB with RCLK connected to ex pin1 of bk1080.. Ready to test bin file once you share 👍

— Reply to this email directly, view it on GitHub https://github.com/phdlee/uvk5cec/issues/6#issuecomment-1908292997, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE4DZPDQEJXFZFSGKWRM2FLYQEOCJAVCNFSM6AAAAABCCJD6GCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMBYGI4TEOJZG4 . You are receiving this because you commented.Message ID: @.***>

[phdlee]

there The email could not be sent continuously due to a file size issue, so I sent it repeatedly.

2.I found out why it didn't work. It seems the clocks are not satisfactory. That's because the clock size is small and the bias is at 0 volt. In uv-k5, the clock comes through cap. It is the cap on the left of pin 1 of the bk1080. As it passes through the cap, the bias becomes 0volt. It seems that bk1080 can operate that way. However, when I experimented with si4732-a10, it was the same as a normal digital clock (1/2 of vcc).

The photo below shows the minimum operating conditions. [image: image.png]

On Thu, Jan 25, 2024 at 12:13 AM Ian Lee @.***> wrote:

2.I found out why it didn't work. It seems the clocks are not satisfactory. That's because the clock size is small and the bias is at 0 volt. In uv-k5, the clock comes through cap. It is the cap on the left of pin 1 of the bk1080. As it passes through the cap, the bias becomes 0volt. It seems that bk1080 can operate that way. However, when I experimented with si4732-a10, it was the same as a normal digital clock (1/2 of vcc).

The photo below shows the minimum operating conditions. [image: image.png]

It started to operate when the above conditions were met. [image: image.png]

On Wed, Jan 24, 2024 at 11:54 PM iu0azb @.***> wrote:

That's great.. At the moment I've installed the si4732 on the K5 PCB with RCLK connected to ex pin1 of bk1080.. Ready to test bin file once you share 👍

— Reply to this email directly, view it on GitHub https://github.com/phdlee/uvk5cec/issues/6#issuecomment-1908292997, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE4DZPDQEJXFZFSGKWRM2FLYQEOCJAVCNFSM6AAAAABCCJD6GCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMBYGI4TEOJZG4 . You are receiving this because you commented.Message ID: @.***>

[phdlee]

When I attach photos or files, an error continues to occur. I'm removing all photos and reposting only the text. If you need firmware for UV-K5 SI4732 for 26mhz OSC, please let me know your personal email address. kd8cec at gmail com I will send it to you all tomorrow. A photo is not attached, but this is firmware that has been confirmed to work with a different oscillator. (26Mhz with 1.5 volt) Below is the original email I wanted to send today. (photo remove

Ian KD8CEC

iu0azb, jh4vaj According to the datasheet, it looks like PC3 can output a clock from a timer. I can't be sure because I can't see the circuit diagram right now (it's somewhere other than my computer), but I'll check for recycling of the SWD port tomorrow. today I found out a few things

1. I worked on sending the firmware to IU0AZB. I had to be careful because I had to send firmware that didn't work in my environment. I looked at the code to make sure the si4732 library was working correctly and found a few unknown bugs. si4732 has two pins that receive clock input, and I took turns testing them. However, as you can see in the library source, that part is commented out. If the 12th BIT is 1, it is DCLK, and if it is 0, it is the REFC pin. It seems to have been commented out because it does not work while shifting 1 to the left 13 times. Anyway, we have checked all reference clock related parts, almost completed the firmware for 26Mhz clock, and completed all checks. I'll send it tomorrow, and I'll send it to everyone who needs it. please send me your email address @.***

image.png

2.I found out why it didn't work. It seems the clocks are not satisfactory. That's because the clock size is small and the bias is at 0 volt. In uv-k5, the clock comes through cap. It is the cap on the left of pin 1 of the bk1080. As it passes through the cap, the bias becomes 0volt. It seems that bk1080 can operate that way. However, when I experimented with si4732-a10, it was the same as a normal digital clock (1/2 of vcc).

The photo below shows the minimum operating conditions. image.png

The operation starts when the above conditions image.png

I tried increasing the bias of the clock, but without success.

The photo below is for the next experiment. Please ignore the MCU pins being removed or the PCB being scribbled lol

image.png

ian KD8CEC

On Wed, Jan 24, 2024 at 11:54 PM iu0azb @.***> wrote:

That's great.. At the moment I've installed the si4732 on the K5 PCB with RCLK connected to ex pin1 of bk1080.. Ready to test bin file once you share 👍

— Reply to this email directly, view it on GitHub https://github.com/phdlee/uvk5cec/issues/6#issuecomment-1908292997, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE4DZPDQEJXFZFSGKWRM2FLYQEOCJAVCNFSM6AAAAABCCJD6GCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMBYGI4TEOJZG4 . You are receiving this because you commented.Message ID: @.***>

[jh4vaj]

I haven't obtained the SI4732 chip yet, so I can't try the firmware at this time. I applaud your work.

[iu0azb]

I wrote an email, but it wasn't all sent due to capacity issues, so I'll send it in several parts. iu0azb, jh4vaj According to the datasheet, it looks like PC3 can output a clock from a timer. I can't be sure because I can't see the circuit diagram right now (it's somewhere other than my computer), but I'll check for recycling of the SWD port tomorrow. today I found out a few things 1. I worked on sending the firmware to IU0AZB. I had to be careful because I had to send firmware that didn't work in my environment. I looked at the code to make sure the si4732 library was working correctly and found a few unknown bugs. si4732 has two pins that receive clock input, and I took turns testing them. However, as you can see in the library source, that part is commented out.. All the DCLK pins I tested were unnecessary lol If the 12th BIT is 1, it is DCLK, and if it is 0, it is the REFC pin. It seems to have been commented out because it does not work while shifting 1 to the left 13 times. Anyway, I have checked all reference clock related parts, almost completed the firmware for 26Mhz clock, and completed all checks. I'll send it tomorrow, (checked firmware bin) and I'll send it to everyone who needs it. please send me your email address @. [image: image.png] … On Wed, Jan 24, 2024 at 11:54 PM iu0azb @.> wrote: That's great.. At the moment I've installed the si4732 on the K5 PCB with RCLK connected to ex pin1 of bk1080.. Ready to test bin file once you share 👍 — Reply to this email directly, view it on GitHub <#6 (comment)>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE4DZPDQEJXFZFSGKWRM2FLYQEOCJAVCNFSM6AAAAABCCJD6GCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMBYGI4TEOJZG4 . You are receiving this because you commented.Message ID: @.***> iu0azb at gmail ... thanks

[phdlee]

I just appreciate your interest.

On Thu, Jan 25, 2024 at 12:49 AM jh4vaj @.***> wrote:

I haven't obtained the SI4732 chip yet, so I can't try the firmware at this time. I applaud your work.

— Reply to this email directly, view it on GitHub https://github.com/phdlee/uvk5cec/issues/6#issuecomment-1908405747, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE4DZPGW6K2NAKURUOOL4NLYQEURRAVCNFSM6AAAAABCCJD6GCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMBYGQYDKNZUG4 . You are receiving this because you commented.Message ID: @.***>

[iu0azb]

Yeap, unfortunately after have installed the si4732 on the PCB defenitly I''ve reconginze that the chip became deaf in HF :( too many welding cycles.. that's bad. For now I'm stuck untill new chips will be here on the desk
Btw, I've also discovered that on my K5 actually the PA15 is always at 0V.. that's probably why it suffered from first day from poor FM reception.

[phdlee]

There is no PA15. buit GPIO B15 ? When B15 is 0, the RF signal is connected. Press the EXIT button on the radio to switch to High.

The firmware I am sending you now has been checked for operation by setting it to 26Mhz with an external oscillator. And si4732 does not break easily. The radio screen will appear even if only the sclk, sdio pin, gnd, and vcc of the si4732 are connected. If the si4732 chip is dead, the radio screen will not appear. The radio will stop while waiting indefinitely for communication. If you press the 0 button to switch to the radio screen, si4732 is normal.

Thanks~

On Thu, Jan 25, 2024 at 6:34 AM iu0azb @.***> wrote:

Yeap, unfortunately after have installed the si4732 on the PCB defenitly I''ve reconginze that the chip became deaf in HF :( too many welding cycles.. that's bad. For now I'm stuck untill new chips will be here on the desk Btw, I've also discovered that on my K5 actually the PA15 is always at 0V.. that's probably why it suffered from first day from poor FM reception.

— Reply to this email directly, view it on GitHub https://github.com/phdlee/uvk5cec/issues/6#issuecomment-1908953675, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE4DZPEWTKKRVRKRERAIQK3YQF5AFAVCNFSM6AAAAABCCJD6GCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMBYHE2TGNRXGU . You are receiving this because you commented.Message ID: @.***>

[iu0azb]

I tried to investigate using the schematic.. where the base of Q03 goes to PA15.. as also reported on my original schematic But when look at MCU schematic, you are right, it doesn't exist.. and B15 is reported as unused. Anyway I've indetified the problem, L56 coil (0.36uH) is open.. I'll replace it.

I''ve removed the si4732 and installed on my test board with arduino to confirm that the AMI input is gone.. not sure how this happen. Also connecting the chip with an external dipole broadcast signals are really weak.

Let's wait then

[phdlee]

Ah, the schematic was wrong. When I include pictures, an error occurs and I can't post, but the two you included were sent successfully. This is what I marked on the CPU for comparison with the source code. [image: image.png]

ian

On Thu, Jan 25, 2024 at 6:05 PM iu0azb @.***> wrote:

I tried to investigate using the schematic.. where the base of Q03 goes to PA15.. as also reported on my original schematic image.png (view on web) https://github.com/phdlee/uvk5cec/assets/63300383/af090537-2c9a-4391-a807-a97fe95193f6 But when look at MCU schematic, you are right, it doesn't exist.. and B15 is reported as unused. image.png (view on web) https://github.com/phdlee/uvk5cec/assets/63300383/afccc9fc-0a2b-4bef-83bc-23465e469f41 Anyway I've indetified the problem, L56 coil (0.36uH) is open.. I'll replace it :)

— Reply to this email directly, view it on GitHub https://github.com/phdlee/uvk5cec/issues/6#issuecomment-1909702761, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE4DZPFNGHSJHGXPRABLNDTYQIN35AVCNFSM6AAAAABCCJD6GCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMBZG4YDENZWGE . You are receiving this because you commented.Message ID: @.***>

[jh4vaj]

The schematic may not be complete. Perhaps PA15 is a typo.

Looking at this photo, it seems that all pins of the MPU are used.

https://github.com/ludwich66/Quansheng_UV-K5_Wiki/blob/main/hardware/UV-K5_1.4_PCB_2024-01-22_1900px/UV-K5_1.4_PCB_4%2BBT_1900.png

Note that this photo appears to be a view from the back. Therefore, it is mirrored left to right.

[iu0azb]

Yeap, the reverse engineering in schematics is not an easy task :) So, no i/o available on MCU, mean no way to reset the si4732 other than with RC. I look forward to receiving the new chips and go ahed.
I've time to rethink how to place the chip in the area. I had instinctively placed it toppled as in picture, but I'm sure it was not the most effective position.

I also realized that exactly the thickness of the si4732 passes between the plastic of the cover and the PCB ... an usefull info fo who want to enjoy the project

[jh4vaj]

I read new post. so, it was possible to omit the C and R for the reset. That's great. I have designed a PCB to mount additional parts (this includes the C and R). Currently, due to Chinese New Year, the PCB manufacturers in China are on holiday, so once that's over, I plan to place an order.

[iu0azb]

I read new post. so, it was possible to omit the C and R for the reset. That's great.

I have designed a PCB to mount additional parts (this includes the C and R).

Currently, due to Chinese New Year, the PCB manufacturers in China are on holiday, so once that's over, I plan to place an order.

The PCB would be the best cleaner way to mod the radio.. I've just the doubt if there will be enough room to place it. Actually on my radio the space between the pcb and speaker is almost exactly the thickness of si4732 🙁

[jh4vaj]

I will order it as Flex-PCB. Its thickness is 0.11mm. I am hoping it will fit in the enclosure. However, this is my first experience with Flex-PCB, so I am worried.

[iu0azb]

Wow never used too ;)

[phdlee]

It's just something I've been thinking about. You did it. That's cool. If it doesn't fit a little, you can replace the speaker, and it works even if you force it in, but there's no major problem. My first version was a bit thick. If it is operated with a PCB, including C and R will be much more convenient for users who modify it. This is because it is much easier to attach the PCB on top than to put a jumper with POINT.A (RESET). I will continue to maintain C and R versions as well. I'm so excited. Even if it doesn't fit well, send me the Gerber file and I will order it and test it. Ian KD8CEC

On Tue, Feb 13, 2024 at 1:41 AM jh4vaj @.***> wrote:

I will order it as Flex-PCB. Its thickness is 0.11mm. I am hoping it will fit in the enclosure. However, this is my first experience with Flex-PCB, so I am worried.

— Reply to this email directly, view it on GitHub https://github.com/phdlee/uvk5cec/issues/6#issuecomment-1939116752, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE4DZPD3SL3PLJYRXI2LR23YTJA2TAVCNFSM6AAAAABCCJD6GCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMZZGEYTMNZVGI . You are receiving this because you commented.Message ID: @.***>

[iu0azb]

Hi Ian, I read your new post about the si4732 experiments.. you did amazing improvements on that. no more RC!! About the clock at this point I will go ahead with the xtal solution.. not big deal compared with the two bias resistor.. It make software easier :) Fortunately I've received the chips and I'm ready to install it on radio.. At this time I'll wait you release the new bin file to avoid RC Thanks again for your great job! Luca

[jh4vaj]

I have placed a 1uF cap for MOD S2 on the PCB. Then, I printed it on paper and tried to put it on the UV-K5's PCB. It seems to fit nicely.

I am thinking of expanding this PCB a little and adding a dummy pad to fix it to the UV-K5's PCB. Once that's done, I will order this PCB.

[iu0azb]

Hi all, after few days of test I would like to share my feedback on this :) If you scroll up I mentioned that once isntalle the si4732 on the radion it became deaf... but later with new chip I recognised that the probem occurred again.. and confirmed on two different K5. Once I connect the AMI to pin 6 or C point with or without the additonal cap the si4732 saturate and no signals can be detected also attaching an external hf antenna. If I connect same antenna on the chip pin 8, insulating RF input from K5, the reception is at normal level ok an si4732.. I've tried to play a bit aronund the Q7 with no success. With oscope I note a strong noise coming from bias circuit that till now I can't filter out. What's your results on this? I read that PA2JSA added a cap to ground on pin8.. that's improve the noise but wouldn't fix the situation :)

[phdlee]

First, don't connect the AMI to anything.

1. Attach a 10cm wire to the FMI and check if FM radio is being received.
2. Please test whether it works well when switching AM/FM/LSB/USB and changing frequencies. If everything is normal up to this point, SI4732-A10 is installed properly. If possible, it would be helpful to understand the situation if you took a video of this operation and uploaded it. 3.If everything went well up to this point, remove the wire attached to the FMI and connect a 10cm wire to the AMI. Connect the other end of the 10cm wire to the antenna. Even if it doesn't work perfectly, you should get acceptable reception. Please note that the antenna's GND must not be connected anywhere.

Also, when testing, do not use an adapter or USB-C cable, but connect it to the battery and test it. The tests so far have not used any of the K5's RF PATHs. If you tell me the test results, I may be able to figure out the problem.

Ian KD8CEC

On Tue, Feb 20, 2024 at 7:06 PM iu0azb @.***> wrote:

Hi all, after few days of test I would like to share my feedback on this :) If you scroll up I mentioned that once isntalle the si4732 on the radion it became deaf... but later with new chip I recognised that the probem occurred again.. and confirmed on two different K5. Once I connect the AMI to pin 6 or C point with or without the additonal cap the si4732 saturate and no signals can be detected also attaching an external hf antenna. If I connect same antenna on the chip pin 8, insulating RF input from K5, the reception is at normal level ok an si4732.. I've tried to play a bit aronund the Q7 with no success. With oscope I note a strong noise coming from bias circuit that till now I can't filter out. What's your results on this? I read that PA2JSA added a cap to ground on pin8.. that's improve the noise but wouldn't fix the situation :)

— Reply to this email directly, view it on GitHub https://github.com/phdlee/uvk5cec/issues/6#issuecomment-1953871545, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE4DZPDIZAZBYSXPTLBMLILYURYTFAVCNFSM6AAAAABCCJD6GCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSNJTHA3TCNJUGU . You are receiving this because you commented.Message ID: @.***>

[iu0azb]

Hi Ian, Thanks to detailed reply. I did the test and attaching a short wire to FMI it received normally.. same connecting it to AMI.. with or without rf ground connected. All settings can be modified and edit .. while using the "Q7" circuit for HF I've the issue mentioned earlier.

[iu0azb]

Short wire connnected only to AMI

https://github.com/phdlee/uvk5cec/assets/63300383/9e14e446-f334-4a8b-a95a-6cb1365c4305

[iu0azb]

Short wire connected to AMI while connected to Q7 circuit.. sorry no dx blocking cap.. but it's the same

https://github.com/phdlee/uvk5cec/assets/63300383/0772fce7-83cd-4184-b21d-ac32d05658aa

[iu0azb]

Same wire connected to the C7 cap pad before Q7 .. AMI connected as it suppose to stay pin6

https://github.com/phdlee/uvk5cec/assets/63300383/b8b822fb-b582-45e7-bc47-78836fe297ae

[phdlee]

Thank you for sending the video. The test I suggested is to connect the antenna directly to the AMI chip of SI4732-A10 as shown below. Connect the antenna directly to AMI on SI4732-A10 and move to a place where there is always a shortwave radio frequency or (USB) signal such as 14.074Mhz. Connecting the antenna directly to the AMI generally creates an environment similar to that of a radio using the SI4732-A10. This should work regardless of UV-K5 characteristics.

Another interesting thing is that when I touch it with my hand, I can hear shortwave broadcasting. If it is a normal 15Mhz shortwave broadcast, the signal seems too big. I think it would be a good idea to try change the frequency. The crystal and CAP connection is so cool.

Ian KD8CEC [image: image.png]

On Wed, Feb 21, 2024 at 9:29 PM iu0azb @.***> wrote:

Same wire connected to the C7 cap pad before Q7 .. AMI connected as it suppose to stay pin6

https://github.com/phdlee/uvk5cec/assets/63300383/b8b822fb-b582-45e7-bc47-78836fe297ae

— Reply to this email directly, view it on GitHub https://github.com/phdlee/uvk5cec/issues/6#issuecomment-1956546975, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE4DZPGFYU2V64DJRTD52U3YUXSEJAVCNFSM6AAAAABCCJD6GCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSNJWGU2DMOJXGU . You are receiving this because you commented.Message ID: @.***>

[iu0azb]

Yeap, on first video the antenna was attached strictly to AMI.. and it works like a charm, Even with 20cm wire I can hear FT8 signals on 20m band :) But as soon I tried to conect with the rest of circuit the issue mask anything..

I forgot to mentioned that I've also verified the signal path and bias on Q7 and noted that Q8 introduce a strong dirty sinewave signal at 176khz 40mVpp.. maybe is not the root couse of issue.. but not sure

[phdlee]

I also thought about another possibility, and the frequency you are receiving frequency may not be a shortwave broadcast. If it is received when you touch it, it may be an FM or AM broadcast.

I've just tested it and found a similar phenomenon to yours in case of FM broadcasts or incorrect broadcasts coming from other bands. (I tried it again, and the characteristic of radio waves coming from FM broadcasting is that they have a large BandWidth.)

Try receiving around 15Mhz to 20Mhz during the day and below 7Mhz at night. I just tested it and was able to receive about 20-40 stations. If tested, connect the wiring as normal and connect the antenna to the antenna terminal of the original UV-K5 to receive signals such as 14.074 or 7.074. Ian KD8CEC

On Wed, Feb 21, 2024 at 9:29 PM iu0azb @.***> wrote:

Same wire connected to the C7 cap pad before Q7 .. AMI connected as it suppose to stay pin6

https://github.com/phdlee/uvk5cec/assets/63300383/b8b822fb-b582-45e7-bc47-78836fe297ae

— Reply to this email directly, view it on GitHub https://github.com/phdlee/uvk5cec/issues/6#issuecomment-1956546975, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE4DZPGFYU2V64DJRTD52U3YUXSEJAVCNFSM6AAAAABCCJD6GCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSNJWGU2DMOJXGU . You are receiving this because you commented.Message ID: @.***>

[phdlee]

Another thing is, could the signal from your antenna be too strong? First, try adjusting the LNA to maximize ATT. What antenna did you use for testing? If this problem occurs because your antenna's signal is too strong and it cannot be controlled with the ATT of the SI4732-A10, you may need to install an external ATT.

Ian KD8CEC

On Wed, Feb 21, 2024 at 11:48 PM Ian Lee @.***> wrote:

I also thought about another possibility, and the frequency you are receiving frequency may not be a shortwave broadcast. If it is received when you touch it, it may be an FM or AM broadcast.

I've just tested it and found a similar phenomenon to yours in case of FM broadcasts or incorrect broadcasts coming from other bands. (I tried it again, and the characteristic of radio waves coming from FM broadcasting is that they have a large BandWidth.)

Try receiving around 15Mhz to 20Mhz during the day and below 7Mhz at night. I just tested it and was able to receive about 20-40 stations. If tested, connect the wiring as normal and connect the antenna to the antenna terminal of the original UV-K5 to receive signals such as 14.074 or 7.074. Ian KD8CEC

On Wed, Feb 21, 2024 at 9:29 PM iu0azb @.***> wrote:

Same wire connected to the C7 cap pad before Q7 .. AMI connected as it suppose to stay pin6

https://github.com/phdlee/uvk5cec/assets/63300383/b8b822fb-b582-45e7-bc47-78836fe297ae

— Reply to this email directly, view it on GitHub https://github.com/phdlee/uvk5cec/issues/6#issuecomment-1956546975, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE4DZPGFYU2V64DJRTD52U3YUXSEJAVCNFSM6AAAAABCCJD6GCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSNJWGU2DMOJXGU . You are receiving this because you commented.Message ID: @.***>

[iu0azb]

I'm sure that chip works.. here a short video connecting a longer wire receiving 40m band. I'd be strongly satisfied if get same results using original SMA path.

https://github.com/phdlee/uvk5cec/assets/63300383/ce2f6e15-912b-4475-8a7d-3e0be1e5b1eb

Once I try to connect AMI to L56 any signals drop down..

[phdlee]

Another way is to skip Q17. You can remove Q17 and bypass it, but when using a portable antenna, you need to check whether reception performance is poor. Even if Q17 is removed, there does not seem to be a major problem with the radio's performance because TX/RX is controlled by the PIN diode.

On Wed, Feb 21, 2024 at 11:45 PM iu0azb @.***> wrote:

Yeap, on first video the antenna was attached strictly to AMI.. and it works like a charm, Even with 20cm wire I can hear FT8 signals on 20m band :) But as soon I tried to conect with the rest of circuit the issue mask anything..

I forgot to mentioned that I've also verified the signal path and bias on Q7 and noted that Q8 introduce a strong dirty sinewave signal at 176khz 40mVpp.. maybe is not the root couse of issue.. but not sure

image.png (view on web) https://github.com/phdlee/uvk5cec/assets/63300383/cbce75d9-a598-4cfb-87a1-25381c149229

— Reply to this email directly, view it on GitHub https://github.com/phdlee/uvk5cec/issues/6#issuecomment-1956809329, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE4DZPHAJUP3VA4ALTDZ5X3YUYB7XAVCNFSM6AAAAABCCJD6GCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSNJWHAYDSMZSHE . You are receiving this because you commented.Message ID: @.***>

[iu0azb]

https://github.com/phdlee/uvk5cec/assets/63300383/3896cd93-734a-48e3-be94-0d8fe95f757e

[phdlee]

If you can test it, would you try bypassing Q17 as shown below? Connect the antenna to the UV-K5 antenna connector.

Or, how about reattaching the removed C102 (connected and disconnected in the last video) or connecting it with a capacitor of about 1uF? (At this time, the antenna must be connected to the antenna connector of UV-K5) [image: image.png]

On Thu, Feb 22, 2024 at 12:06 AM iu0azb @.***> wrote:

https://github.com/phdlee/uvk5cec/assets/63300383/3896cd93-734a-48e3-be94-0d8fe95f757e

— Reply to this email directly, view it on GitHub https://github.com/phdlee/uvk5cec/issues/6#issuecomment-1956868120, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE4DZPE7SW4YM2F7B3D4ZGTYUYENVAVCNFSM6AAAAABCCJD6GCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSNJWHA3DQMJSGA . You are receiving this because you commented.Message ID: @.***>

[iu0azb]

Bypassing Q17 I got similar issue, no strong noise as when it's active. C102 is having no impact.. initially, following your reccomandation I've inserted 1uF.. but investigating I've removed.

I don't want to disturb you further... I assume that the problem is not common and therefore not of wide interest..

Anyway I just tried to connect AMI directly to the input of C99.. and the signal drop down. I therefore need to understand the path of the FM_Radio_Ant to move forward... there could be something on the other side of PCB that impacting..

I can replicate the problem on both uvk5 I've .. that's really wiird

BTW, I forgot to mention earlier that I noted only on AM reception the BW changes apparently no act.. not sure if is an know bug or not :)

https://github.com/phdlee/uvk5cec/assets/63300383/c6f69c6e-e792-4478-a531-432b2e88ab5f

[phdlee]

Thank you for sending us the various experiment results.

Please try the below when you have time.You don't have to reply. Lastly, how about connecting everything normally and connecting one of the antenna wires directly to the UV-K5's antenna terminal? (Excluding the antenna ground, only the antenna center line is connected directly to the center of the UV-K5 connector)

The UV-K5's antenna connector is not standard, so all my cables and contacts are bad. (90% of the cables I have do not fit well.) So, when shooting video, I only use the cable I always use.

Thank you for your hard work testing. Please let me know if you have any other results later. Ian KD8CEC

On Thu, Feb 22, 2024 at 12:23 AM iu0azb @.***> wrote:

Bypassing Q17 I got similar issue, no strong noise as when it's active. C102 is having no impact.. initially, following your reccomandation I've inserted 1uF.. but investigating I've removed.

I don't want to disturb you further... I assume that the problem is not common and therefore not of wide interest..

Anyway I just tried connecting AMI directly to the input of C99.. and the signal drop down. I therefore need to understand the path of the FM_Radio_Ant to move forward... there could be something on the other side of PCB that impacting..

— Reply to this email directly, view it on GitHub https://github.com/phdlee/uvk5cec/issues/6#issuecomment-1956932629, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE4DZPENUO57Z33MINODWWLYUYGQNAVCNFSM6AAAAABCCJD6GCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSNJWHEZTENRSHE . You are receiving this because you commented.Message ID: @.***>

[iu0azb]

At the end of a day spent to understand the problem I decide to bypass it, adding a dedicated SMA in the LED spot 😊. For who want to try, be informed that there is almost no space between case and pcb/lcd flex. The back of SMA need to be plunged inside plastic and back need to be modified to void any extra space other than central wire.

Cool working result.

[alboc2k]

Hello Ian, I made the mod and it works quite well. If you like, I can send you some videos and photos. It dosen't look pretty, but it works. Probably I'll go for a sma port like Luca did.

Things to mention/suggestions:

1. add some sort of memory for Commercial FM and for band VFO
2. More use of keyboard instead of menu+side buttons.
3. Like said by Luca there are some noise/click, so maybe a sma port should be better.
4. With (hi gain) antennas loks like there is some sort of saturation. Again, maybe point 3 resolves
5. a notch for FM or (better) a bandpass filter is mandatory with hig gain antennas

Thank you for your wonderful work!

73 de IU3POA.

[iu0azb]

Someone was too fast then us 😳

https://a.aliexpress.com/_Ey5oFHF