IIRC the ESP32 processor is similar in "power" to the CortexM0+, so I would image that the dual core rp2040 processor (pico) should also be usable here.
I belive that rp2040 won't be able to hande the Hilbert transformer FIR as shown in this video. To handle the load, the ESP32 processor uses (clocks/sec) = 351taps x 4mac x 44.1kHz x 2channels = 124MHz, which is doable on 160MHs ESP core. Here are only 4 clocks are used to perform floating point multiply - accumulate (MAC) using assembler based DSP library, provided by the manufacturer. The rp2040 has neither floating point arithmetic, nor assembly-based DSP library, Even on faster M4 core, single MAC takes 7 clocks. So bottom line, it is unlikely to fit the rp2040 without compromises.
I haven't played around with the rp2040 at all I did do some playing around with DSP on the STM F7 uController. It was single core 160Mhz and it could do about 160 taps and there was little horsepower left over to do anything else. That was using CMSIS DSP and the F7 did have a math processor. Its a shame as I did like STM's tooling.
@@na5y just wonder, where do you get the time to play with electronics and to make a video. Great stuff!. Thank you for the detailed explanations and quality of the material.
Great amount of stuff to learn here - where did you pick up the EN8388 module? I only find the chips (probably clones) at the regular eastern suppliers...
The ES8388 module is from pcbartists and can be found here: pcbartists.com/product/es8388-module/ I plan on using this combo also as it has an ESP32 and ES8388 in the one package: www.aliexpress.us/item/3256803619642039.html?spm=a2g0o.productlist.0.0.23606c91rbHJPW&algo_pvid=ad742fa2-b0e8-4044-9159-54c56c1189cf&algo_exp_id=ad742fa2-b0e8-4044-9159-54c56c1189cf-0&pdp_ext_f=%7B%22sku_id%22%3A%2212000027228845836%22%7D&pdp_npi=2%40dis%21USD%213.86%213.74%21%21%21%21%21%402101e9d116708760879394577e95ee%2112000027228845836%21sea&curPageLogUid=zlZGgHlTEeWT It doesn't have pin headers though or a USB interface
@@na5y Thanks a lot the ali one is a great find - I'm going to order 2 of them to follow along as you build - really learning a lot from your channel over the last few days! I'm not afraid of not having pins :)
Nice work Ian. Do you know what the value of the Microphone PGA gain in Register 9 ADC Control 1 is. It can be set for +24dB, just curious why you needed a microphone preamp. If you mention in the video my apology. Thanks Video is awesome can't wait for the next. 73
I have honestly never tried that. Would be great to not have to have the preamp. You'd still have to have power to the electret of course. I might give that a try - thanks for the suggestion and comment!
@@na5y Yes of course you still need to bias the microphone. I found a quick way to change the PGA is in ES8388.cpp in the ".init" method. The data sheet recommends that you use balance mode but just for a test should be fine. Thanks again Ian
I’m enjoying your SDR videos 😀
Thank you Julias - I hope to complete it one day...
Nice vlog. Question why dont you reduce the audio sampling freq from 44k to say 8k so then you wont need a software low pass filter of 4kHz?.
Thank you for the comment. There's no real reason to use 44kHz, frankly and it would certainly make the processing requirements easier.
Great Video , Nice , Waiting For Complete Transceiver..
Coming soon!
IIRC the ESP32 processor is similar in "power" to the CortexM0+, so I would image that the dual core rp2040 processor (pico) should also be usable here.
Potentially - assuming CMSIS-DSP has an appropriately tuned FIR algorithm for that model. I think it does. Thank you for the comment!
I belive that rp2040 won't be able to hande the Hilbert transformer FIR as shown in this video. To handle the load, the ESP32 processor uses (clocks/sec) = 351taps x 4mac x 44.1kHz x 2channels = 124MHz, which is doable on 160MHs ESP core. Here are only 4 clocks are used to perform floating point multiply - accumulate (MAC) using assembler based DSP library, provided by the manufacturer. The rp2040 has neither floating point arithmetic, nor assembly-based DSP library, Even on faster M4 core, single MAC takes 7 clocks. So bottom line, it is unlikely to fit the rp2040 without compromises.
I haven't played around with the rp2040 at all I did do some playing around with DSP on the STM F7 uController. It was single core 160Mhz and it could do about 160 taps and there was little horsepower left over to do anything else. That was using CMSIS DSP and the F7 did have a math processor. Its a shame as I did like STM's tooling.
@@na5y just wonder, where do you get the time to play with electronics and to make a video. Great stuff!. Thank you for the detailed explanations and quality of the material.
Well being currently retired does help a bit ;)
great video, i look forward to the next one
Thank you Khim!
Great Job, I like it a lot!!
Thank you - should have the next in the series out soon
@@na5y Looking forward to watch it Ian!!! Thank you for all your excellent work!!1
Thanks, Ian!
Glad you enjoyed Alexandr!
Great amount of stuff to learn here - where did you pick up the EN8388 module? I only find the chips (probably clones) at the regular eastern suppliers...
The ES8388 module is from pcbartists and can be found here:
pcbartists.com/product/es8388-module/
I plan on using this combo also as it has an ESP32 and ES8388 in the one package:
www.aliexpress.us/item/3256803619642039.html?spm=a2g0o.productlist.0.0.23606c91rbHJPW&algo_pvid=ad742fa2-b0e8-4044-9159-54c56c1189cf&algo_exp_id=ad742fa2-b0e8-4044-9159-54c56c1189cf-0&pdp_ext_f=%7B%22sku_id%22%3A%2212000027228845836%22%7D&pdp_npi=2%40dis%21USD%213.86%213.74%21%21%21%21%21%402101e9d116708760879394577e95ee%2112000027228845836%21sea&curPageLogUid=zlZGgHlTEeWT
It doesn't have pin headers though or a USB interface
@@na5y Thanks a lot the ali one is a great find - I'm going to order 2 of them to follow along as you build - really learning a lot from your channel over the last few days! I'm not afraid of not having pins :)
great explanation Ian, keep them coming 🎉🎉🎉🎉🎉🎉
Thank you Joe!
Nice work Ian. Do you know what the value of the Microphone PGA gain in Register 9 ADC Control 1 is. It can be set for +24dB, just curious why you needed a microphone preamp. If you mention in the video my apology. Thanks Video is awesome can't wait for the next. 73
I have honestly never tried that. Would be great to not have to have the preamp. You'd still have to have power to the electret of course. I might give that a try - thanks for the suggestion and comment!
Just checked in es8388.cpp its hardcoded to 0x11 - +3dB per channel
@@na5y Yes of course you still need to bias the microphone. I found a quick way to change the PGA is in ES8388.cpp in the ".init" method. The data sheet recommends that you use balance mode but just for a test should be fine. Thanks again Ian
🔥👍
Thanks Alex!
G'day Alex!
@@Paul_VK3HN 👋