[0:03] my new esp32 TV boards have arrived from
[0:06] PCB way there are a couple of problems
[0:08] with the board which we’ll cover next
[0:10] but there’s also a catastrophic error on
[0:12] the board that could have led to some
[0:14] really nasty consequences so quick quiz
[0:17] can you see the problem is it a the
[0:19] moset is the wrong way around B the
[0:22] input capacitor is way too low C the
[0:25] ncp1 167 can’t deliver enough power or D
[0:29] something else that really obvious we’ll
[0:31] come back to this towards the end of the
[0:32] video let me know in the comments if you
[0:34] get it
[0:36] right my goal with this board was to get
[0:39] everything working we’ve got audio
[0:41] output from the max
[0:42] 98537 battery charging with a tp4056 an
[0:46] infrared receiver and the biggest
[0:48] feature of all super high-speed access
[0:50] to the SD card now this is the most
[0:53] exciting thing on the board if it works
[0:55] then we can load files onto our TV
[0:57] without needing to unplug the SD card
[1:00] now it is already possible to access the
[1:01] SD card directly from our PC using the
[1:04] esp32 in mass storage mode but it’s
[1:06] painfully slow I did manage to get it up
[1:09] to 1 Megabyte per second reading and
[1:11] writing but it took a lot of hacking and
[1:13] for me it’s still not fast enough
[1:15] especially if you want to transfer large
[1:17] video files so that’s the big goal
[1:20] high-speed SD card access over USB to
[1:23] achieve this I’ve added an additional IC
[1:26] the USB 2244 in theory this should give
[1:29] us a transfer rate to the SD card of up
[1:31] to 35 mbes per second now this does of
[1:35] course depend on the SD card and how
[1:37] good our USB connection is testing the
[1:39] SD card I’m using connected directly to
[1:41] my PC I get 25 megabytes per second so
[1:44] that’s my target rate if we can get
[1:46] close to that I’m going to call it a
[1:50] win so the big question is does the new
[1:53] PCB actually work do we get the results
[1:56] I’m looking for and initially it was a
[1:59] big fat no the board did absolutely
[2:02] nothing we can’t connect to the esp32
[2:05] and we can’t connect to the SD card what
[2:07] a disappointment we’re completely dead
[2:10] when I made the board I was concerned
[2:12] that the es32 and the USB 2244 would not
[2:15] play nicely together but I was hoping
[2:18] that the es32 would be able to use the
[2:20] USB connection when the USB 2244 was
[2:23] held in reset mode to ensure this I put
[2:26] a pull down resistor on the reset line
[2:28] when it’s in reset mode all the pins are
[2:30] supposed to go into a high impedance
[2:32] State and not interfere with anything
[2:34] else looking at my schematic I can see
[2:36] that I’d Ed gpio 39 to control the reset
[2:39] line as we discovered in our previous
[2:42] video printing pcbs at home this pin
[2:44] behaves weirdly when the esp32 is in
[2:47] download mode both gpio 39 and gpio 40
[2:51] are used for JTAG debugging so you get
[2:54] all sorts of weird output on them my
[2:56] initial suspicion is that this was
[2:58] stopping the USB 224 4 from going into
[3:00] reset and that it was conflicting with
[3:02] the es32 on the USB lines measuring the
[3:06] voltage on the reset pin I got 2.7 volts
[3:09] to me this felt like some kind of clock
[3:11] signal or a data line that is being
[3:13] toggled on and off so I attacked my PCB
[3:16] and chopped through the trace leaving
[3:18] just the pull down resistor the USB 2244
[3:21] should now definitely be in reset mode
[3:24] all the time sadly this did nothing I
[3:27] still couldn’t connect to the ESB 32
[3:30] this completely stumped me so I took the
[3:32] nuclear option and Des soed the USB 2244
[3:35] completely leaving just the ESB 32
[3:38] module and as expected I could now
[3:40] connect to the esp32 without any
[3:43] problems so I tried the opposite I took
[3:46] another board and desol the USB 32
[3:49] plugging this in I now got an SD card
[3:51] appearing on my computer which was great
[3:54] the USB 2244 was working but also not
[3:57] great the USB 2244 should be held in
[4:00] reset mode VI the pull down resistor it
[4:02] should not start up at all measuring the
[4:05] voltage on the reset pin I still got 2.7
[4:08] volts what on Earth is going on there’s
[4:11] nothing in the data sheet but looking at
[4:13] some other similar ic’s I think there is
[4:15] an undocumented pull-up resistor on the
[4:17] reset line giving our measured voltage
[4:20] of 2.7 volts and the 100K pull down
[4:23] resistor this hidden pull-up resistor
[4:25] must be around 20K our 100K pull down
[4:28] resistor is not going to do anything so
[4:31] I took another board thank goodness I
[4:33] ordered five assembled boards and I
[4:35] solded a 1K resistor across the existing
[4:37] pull down resistor and hey Presto the
[4:40] esp32 sprang into life I then tested out
[4:43] switching over from the esp32 to the USB
[4:47] 2244 I changed the two USB pins on the
[4:50] USB 32 to input and enabled the USB 2244
[4:53] by taking gp39 high and amazingly my
[4:57] board transformed into a high-speed SD
[4:59] card my proof of concept works we can
[5:02] share the USB bus between the two chips
[5:04] and switch between them all it took is
[5:07] one bodge resistor but just how fast is
[5:10] our high-speed SD card with the USB 2244
[5:14] enabled we can transfer files at around
[5:16] 21 megabytes per second that’s over 20
[5:20] times faster than using the esp32
[5:22] directly it’s a massive Improvement once
[5:25] more we’re close to what I was able to
[5:27] get when connecting the card to my Mac
[5:29] so I’m happy with this performance it’s
[5:31] an amazing result but what about the
[5:34] rest of the board do all the other
[5:36] peripherals work the first thing I
[5:38] wanted to check was the display on the
[5:40] back of the board I’ve got a footprint
[5:42] for the FBC ribbon cable the display has
[5:44] a completely nonstandard pin pitch of 7
[5:47] the only place that I’ve been able to
[5:48] find a connector with this pitch is on
[5:50] Alibaba and I’m not sure I really need
[5:53] one 100 of them at least not yet so
[5:56] we’re going to have to sold it on I’ve
[5:57] never solded one of these flat cables
[5:59] before but it seemed to work okay I just
[6:01] tinned the pads added plenty of flux and
[6:04] then taped the cable in place with
[6:05] capton tape this technique seemed to
[6:07] work okay I kept the soldier iron on
[6:09] quite a low temperature and just run it
[6:11] over each pad until the soldier flowed
[6:14] with my TV firmware loaded and adjusted
[6:16] for the new pins we get a picture
[6:18] showing up and with a speaker plugged in
[6:20] we get sound coming out the amplifier is
[6:22] all
[6:23] [Music]
[6:25] [Applause]
[6:27] good my new esp32
[6:37] TV come
[6:40] [Music]
[6:42] in once more this proves that the
[6:44] infrared receiver is also working we’ve
[6:47] come quite a long way for our original
[6:49] prototype with our minimal Dev board
[6:51] project so what’s next there are a
[6:54] couple of things that need improving I’m
[6:56] very pleased with the sharing of the SD
[6:58] card between the USB 2244 and the ESP 32
[7:02] this seems to work nicely it is possible
[7:05] to turn off the jtac pins by burning an
[7:07] euse on the es32 and holding gpo3 high
[7:11] but of course to do that we need to be
[7:13] able to connect to the es32 a bit of a
[7:16] chicken and egg situation we could use a
[7:19] different Jeep openin to fix this but
[7:21] sharing the USB connection between the
[7:23] two ICS concerns me it just feels a bit
[7:26] nasty and way too easy to get into a
[7:28] state where we can’t talk to the es32
[7:30] and our board is brooked so I’m thinking
[7:33] of moving to having two USB sockets one
[7:35] for the es32 and one for the USB
[7:39] 2244 with the ribbon cable working I can
[7:42] lose this pin header which gives me a
[7:43] bit more space on the board I could even
[7:46] skip the USB socket for the USB 32 and
[7:49] just have a programming header for it
[7:51] another issue I found is that the module
[7:53] I’m using has 8 megab of PS Ram enabling
[7:56] this caused the board to crash as soon
[7:58] as I tried to use the SD card the
[8:00] problem is if we’re using psram we’re
[8:02] not allowed to use gp35 gp36 or GP 37
[8:08] and of course I used all three of these
[8:09] pins for my SD card connection so I’ll
[8:12] have to shuffle a few things around to
[8:13] avoid these pins and the JTAG pins I’m
[8:17] also taking a look at a much cheaper IC
[8:19] for the SD card controller there’s a GL
[8:21] 823 that is just 40 cents that’s 10
[8:25] times cheaper than the USB 2244 and it
[8:29] doesn’t need a crystal which simplifies
[8:30] the circuit and reduces the bomb nicely
[8:33] the docks for this IC are not great so
[8:35] I’m going to create a test PCB first to
[8:37] see if it actually works now I’m sure
[8:40] you’re asking what’s the big issue you
[8:42] promised us a potentially explosive
[8:44] error well I’m not sure where this ever
[8:46] crept in and it’s a bit of a silly one I
[8:49] got it right in the previous board but
[8:51] somehow this diode got move to the wrong
[8:54] side of this mosfet this power control
[8:56] circuit is designed to switch between
[8:57] the USB power and battery power when the
[9:00] USB is connected the battery should be
[9:02] blocked and when the USB is not
[9:04] connected the battery should be used
[9:06] it’s a very cleever circuit when it’s
[9:08] done properly with the current PCB I’ve
[9:11] connected the USB 5 volts directly to
[9:13] the battery via this diode this is
[9:16] definitely not what you’re supposed to
[9:17] do and it could have ended up in a very
[9:19] nasty situation it’s a good job I didn’t
[9:21] test out the battery charging first
[9:23] before I notice this so I think that’s
[9:26] probably the first thing to fix before I
[9:27] do anything else while I working on that
[9:30] you might enjoy this playlist of videos
[9:32] you can see the journey we took to get
[9:33] to this video and what we might be doing
[9:35] next