800x600 looks good to me in Windows 3.11. USB and network are also working as expected. Didn't check the internals yet.

Forgot to mention that I needed some pliers to get the CF card out. It's pretty flush and tight so only using my nails was not enough. 😀 The BIOS issue seems not to be completely gone then. It forgot the date/time 2 times in the beginning. After that it was good for a couple of hours (including soft reset, turn on/off, charging on/off etc multiple times). Will check tomorrow if it's still okay.

The schematic and PCB file show a separate CMOS battery, like a CR2032 holder or similar just tucked under the soundcard... You'd hope its supplied with a fresh battery (and you'd expect that its still in correctly after transit), but I guess its worth checking if you are in there at some point...
the cct used for swapping over from the CMOS battery to the +5V is different from the pocket386, not sure why the designer has done this, only thing I can think of ATM is that potentially the BJT cct has a slightly lower forward drop than a schottky diode , but I cant see why this would be needed! If 3V ish from a battery is enough, that diode drop from 5V shouldn't matter.... so I'm not sure Ive figured that one out yet!

My pocket386 CF slot has got a little less stiff over time! even with Ethernet, it makes a very good sneakernet...

As another aside, part of the PSU redesign is the main battery is now a couple of 18650 cells. So its voltage is prob too high for it to be used directly as the CMOS battery.
From the cct diagrams the main battery charge controller IC seems to be set up for 2 18650 cells in series, for those wanting more battery capacity, the IC can be set to work with 3 cells, space permitting & if you want to add some kind of charge leveling...
I think the extra battery voltage should be OK; the main +5v buck converter runs of the 12V or the battery , the extra voltage might even help with the issue of unplugging the 12V supply with a low battery...

Got mine. Don't have time to play though.
First impression is that it is big (albeit the "like 2 x Pocket386 stacked together" is spot on) , fan is always on with charger connected as far as I can see so also noisy.

Exactly same BIOS version as pocket386 , maybe with changed defaults (mouse support is enabled, didn't see other settings). The BIOS no longer complains of FDC error on boot even if enabled and there's no drive. I didn't get any CMOS reset (yet) but even if we get them they shall be slightly less annoying.

"Touchpad" is also an LCD display. Missed that entirely on the schematic 😀 (And sorry for ruining the jump-scare if you hadn't noticed it either 😁 ) It shows status as what was coming from to the 8051 to the OSD before (also incl. keyboard LEDs), but I'm not sure who is driving this display.
Something is really off about the feeling on the touchpad, maybe it's the small size. But usable. Seems to be using Logitech Mouseman driver on Win side, COM2, no driver installed on DOS side.

The CF slot is harder than on pocket386 as described. The CF sits a bit less flush than before, but entire socket is much more recessed in the chassis, so now I definitely need a tool at least.

CH375 USB driver got newer "2.2" version and is now called XTUSB.SYS. No functional changes noticed yet.

Covox included, came with test program and seems to work.

I used VBIOS (claims to be "GL-GD5420" ) to switch to 1024x768x16colors and while VGA says it worked , panel went blank and OSD displayed "No support". Max mode I can set through BIOS which really worked is 800x600 256 color. Since the connection is still analog and there's some scaling going on no matter what, it's still a bit fuzzy at 800x600. I can't really figure out what is the real resolution of the panel, sorry. But it seems the internals can't go more than 800x600 again.

Thanks for the update.
Mine has just reached UK customs... it might be with me in a few days...
Shame about the fan, I thought you'd be able to turn it off using the OSD? THe manual seem to indicate that i think?
Its a little odd in that its driven from the "IPSOUT" on the schematic, which is the input to the main +5V buck convertor, so approx +12V with the supply plugged in, approx 6 to 8V or so when running off battery. Most 12V fans are guaranteed to run at 6V, but obv they do run slowly, could it just be that? That its going up to full speed when you plug in the charger?
I could be tempted to put a fan controller PCB with a thermistor... there must be an off the shelf one that would do the job, if not its not a big job to design one.... Given the POcket386 is more or less OK without a fan, it does seem a shame to leave it running, esp at full whack if its not needed.

No I hadn't spotted that the touchpad is also a display! Ill call that a bonus! 😉

If its fuzzy at 800x600 it does make me wonder its converting down to 640x480.. Yes, of course you are right it goes through an analogue VGA stage before it goes into a Realtek RTD2280DLW then on to the LCD, but Id have thought if its sampling the analogue signal well & its at the native screen resolution it should give a sharp image??
I could be wrong though, Ive not gone through the datasheet for that IC, for sure it will be using some of its resampling/filtering/dithering etc when its converting resolutions... But when its input is the native screen res Id hope its just "straight through" & expect it to be sharp...
If text looks fuzzy at 800x600 Ill bet the screen is 640x480...
Edit - Incidentally, the pocket 386, when running win95, will let you set the display 800x600 (I dont mean a desktop that auto pans), but it does look truly awful with the rescanning to the 640x480 of the 4:3 mode, not quite as bad at 16:9'

Jules_nerd wrote on 2026-08-05, 18:40:

Shame about the fan, I thought you'd be able to turn it off using the OSD? THe manual seem to indicate that i think?

I did not try any OSD infra whatsoever. I need to RTFM 😀 The old FN buttons still work (albeit FN+F5 shows less info).

Jules_nerd wrote on 2026-08-05, 18:40:

If its fuzzy at 800x600 it does make me wonder its converting down to 640x480..

It is fuzzy at 800x600 but not THAT much fuzzy, I meant to say I cannot really be 100% sure it's not higher than 800x600, but most likely it is 800x600 (just slightly out of sync). Definitely not 640x480 as that would be easily noticeable.

ah OK, The pocket 386 has some settings on the OSD for things like sample clock phase, if its not 'that' fuzzy , and the Book386 also has them, it might be worth a play...
Ahh OK, I get you now! ... youre saying it might be 1024x768, Im not sure how the Realtek RTD2280DLW 'knows' what the resolution of the display is.... presumably some communications between the two goes on for things like that?...
When you say not sure what is driving the touchpad display, now you've pointed it out, it does looks like there are connections to the 8051... (on the cct diagrams I think its connector P4 @ B5 on the CPU sheet)

My mistake, the fan control isnt on the OSD its just an Fn+4:

Fn + F4: Opens the screen's OSD Menu to adjust display parameters such as brightness.
Fn + ~ : Disable the built-in keyboard. After disabling it (the Touchpad will display "KBD OFF"),
an external PS/2 keyboard can be safely inserted and used.
Fn + 1 : VGA OUT / LCD Toggle: Switches the active video output channel between the built-in
LCD screen and the external VGA OUT port.
Fn + 2 : VGA IN (ON/OFF): When activated, the LCD displays "no signal" while waiting for an
external video signal input. Press it again to return to the host system.
Fn + 3 : Touchpad (ON/OFF): Manually enables or disables the built-in touchpad (see "Common
Problems" for usage rules)
Fn + 4 : Fan Switch: Manually turns the internal active cooling fan on or off.
Fn + 5 : 12V/-12V/-5V Boost circuit ON/OFF
Fn + 6 : Enables or disables Covox registers.
Fn + 7 : TFT Switch: Quickly turns the Touchpad TFT backlight power on or off.
Fn + Tab : Power Information Page Switch: TFT showing real-time battery voltage, power status,
and current analytics

I hadn't spotted that you can turn off the =12v/-12V, -5V ccts from the keyboard...
Fn TAB could be interesting too

Some info on the RTD2280DLW:
The native screen resolution and how it handles different input resolutions/timings is stored in tables in an SPI flash, the P25Q40H on the graphics card...
This includes what resolutions it displays treats as not supported/available.
Not sure if it can be forced to treat 1024x768as not available if its the screen resolution!! it would be a bit of an odd thing to do...
Of course then there is whats in the GD5420 VBIOS...

I do think it is 800x600, it is just a bit blurrier overall at 800x600 than the pocket386 was at 640x480, no matter how much I try to auto-adjust. The fonts are not 100% pixel-aligned, so it loses a bit of the charm. May be my impression, needs more tinkering, maybe with some manual adjustment I'll be able to make the pixel grids "match". Color is better on the Book.

I have also already got a CMOS reset once, so it's still there, just seems less frequent (and less annoying because it won't do the FDC error beep).
Sadly I think there's something wrong with the touchpad because it fails to be detected about 2/3 of the time. Even the stock Win3.11 install had this problem, but also MSMOUSE on DOS, Win95, and even CTMOUSE itself will only detect it 1 out of 3 attempts. It's been a while since I last used a serial mouse, but I don't remember them being this unreliable 😀. Strange thing, once connected, it doesn't seem to be dropping packets like crazy, so I'm not sure what is going on. Need someone else to check so that I can discard a hw issue from my unit 😀

On the new display you can monitor power usage on the +5V, which is very interesting. Idle current without coprocessor is about 1.1A , 1.5A with coprocessor ( I have a IIT one rated for 40Mhz ). But, setting the Power Clock Stop bit as per the m6117d datasheet reduces this to 1A (still idling). I had tried this on the p386 and didn't seem to have any effect on battery life or heat, so I forgot about it, but seeing the actual impact here I think makes it worth to do a small utility to set this bit. (Not sure what the con is, since DOS seems to work as normal with the bit set).
Battery voltage is around 8V @85% charge.

Don't have any problems with 800x600 in Windows 3.11. It's very sharp. But I don't have a pocket386 so can't compare.

I mailed the creator about the power adapter and he said they will only ship US plugs worldwide. He said he forgot to add it on Tindie.

Also had a couple of time that it forgot the time/date again. Can't quitte figure out what causes it but I think it's because of the (small) power button. Sometimes I press it not far enough to turn it on. So it quickly turns on/off. Maybe I will replace it in the future as I don't like the power button.

javispedro1 wrote on 2026-08-06, 19:55:

Sadly I think there's something wrong with the touchpad because it fails to be detected about 2/3 of the time. Even the stock Win3.11 install had this problem, but also MSMOUSE on DOS, Win95, and even CTMOUSE itself will only detect it 1 out of 3 attempts. It's been a while since I last used a serial mouse, but I don't remember them being this unreliable 😀. Strange thing, once connected, it doesn't seem to be dropping packets like crazy, so I'm not sure what is going on. Need someone else to check so that I can discard a hw issue from my unit 😀

I tested around 6 mouse programs and half of them had problems detecting it. In the end I settled on the Microsoft mouse version 6.24 driver. It's stable for me for both the touchpad mouse and a serial mouse I tested.

@ javispedro1, OK, maybe manually adjusting the clock phase will sort the slightly blurred screen... IIRC I did adjust that on the pocket...
Its a real shame you're seeing some of these issues, i was really hoping that the pocket's issues would be fully sorted in this model &that it would be more solid... still some of the extra features are nice & it sounds a little more solid than the pocket.
Ill have a look and see if mine has the trackpad issue, it looks like its cleared UK customs so should be here soon, fingers crossed.. OR maybe Ill just install that microsoft driver 😀

On the BIOS/reset front, putting a proper power monitor IC ( PWG node)in one of my pockets totally nailed the BIOS reset issue, the only downside was it kept resetting because of dips in the +5V rail (like for example when you plug/unplug the external PSU or when loud sounds were played), however, on that particular pocket 386 there is definitely a fault that is loading the power supply more than normal, I cant remember how high the current got, but battery life was down from circa 2 hours to 45 mins..... It definitely has a faulty 8051, there should be a replacement IC in my book386 shipment when it arrives , so I can get back on that one (last time, by mistake, they sent me some of the power MOSFETs used to turn the battery ON/OFF & Ive also been sidetracked for a few months playing with some 68k classic macs).. Ill replace that IC and take it from there, but I suspect there will also be another fault somewhere that's drawing too much current.

It did strike me as odd that the book circuit diagrams show a 10uF cap on the output of the power on reset IC (max809), this isnt standard for this IC, in fact there is nothing in the datasheet to say it can safely source/sink the current peaks required to charge/discharge it. Also only the max803, not the 809 needs a pull up resistor??? perhaps this part of the cct diagram is just not current? or maybe some of the parts aren't fitted, perhaps the old RC circuit is still there so they can choose which option to go for when they populate the board???
If all the parts are fitted, I suspect the cap may have been added to help deal with short reset pulses from momentary drops in the +5V. Or the cap could be there to add more delay during switch on. I dont think its great having a slow rising signal on the power good input though, unless there is a Schmitt trigger on it??

Ive certainly seen old reset ICs that can be set to ignore short drops in the supply, they may be worth a go EDIT: on the pocket , or perhaps a lower threshold is the only way...

One difference is that on the reset chip i used, a BD5245-2M, you can set how long a delay after the power supply hits the threshold you have before it sets PWG highEDIT actually the power supply has to remain above the threshold for the full delay time, or it resets the timer (its preset to 140mS on the max80*), I set this to around 1 sec to get around not just switch bounce, but any 'fumbling' with pressing the power switch - Im with you in that the switch is not great really, but with this PWG delay it was fine.
I think there may well also be PSU hold up time issues (at least there is on the pocket) when plugging/unplugging the external power...

Yoghoo wrote on 2026-08-06, 21:35:

I mailed the creator about the power adapter and he said they will only ship US plugs worldwide. He said he forgot to add it on Tindie.

Is it a universal input voltage supply?

Jules_nerd wrote on 2026-08-06, 21:50:

Yoghoo wrote on 2026-08-06, 21:35:

I mailed the creator about the power adapter and he said they will only ship US plugs worldwide. He said he forgot to add it on Tindie.

Is it a universal input voltage supply?

The input is universal but the plug is US. I'm using an US to EU plug converter atm. Ordered an USB-C to DC cable (which will arrive tomorrow) so I can hopefully just use a USB-C charger.

Yoghoo wrote on 2026-08-06, 22:27:

The input is universal but the plug is US. I'm using an US to EU plug converter atm. Ordered an USB-C to DC cable (which will arrive tomorrow) so I can hopefully just use a USB-C charger.

Yes you can, the plug is the mega common 12V barrel one, you can use almost anything, I have never used the original supply on either this or the p386 😀

Yoghoo wrote on 2026-08-06, 21:39:

I tested around 6 mouse programs and half of them had problems detecting it. In the end I settled on the Microsoft mouse version 6.24 driver. It's stable for me for both the touchpad mouse and a serial mouse I tested.

The problem is other OSes. I'm sure I could probably workaround this in ctmouse or mine but whatever the cause is, it will not be really fixable (cause I'm quite sure it will be in the 8051 fw).

As promised here is a small tool to set the "Power Clock Stop" bit as discussed in the SoC datasheet.

Attached both binary and (Open)Watcom source (10KB binary for running what is basically two CPU instructions, sorry).
For using the binary , run "pwrcr on" in real mode (this means, no EMM386 loaded).
Me I'm getting about 100mA less on idle when on a HLT loop (so on DOS you need something like POWER.EXE loaded) .
Disclaimer: I have no idea if this will cause the CPU to catch fire with either short-term or longer-term use. This is just for experimenting and MOST LIKELY will have zero impact on your battery life (unless you leave the device idling/HLTd for significant periods of time).

The USB-C to DC cable works indeed perfectly fine with the BOOK386. So can use an USB charger now.

Also designed a cover for the coprocessor/led gap. Used some clear translucent PETG so the LEDs are still visible. Ordered some darker translucent PETG so it fits the laptop better.

javispedro1 wrote on 2026-08-06, 23:02:

As promised here is a small tool to set the "Power Clock Stop" bit as discussed in the SoC datasheet. […]
Show full quote

As promised here is a small tool to set the "Power Clock Stop" bit as discussed in the SoC datasheet.

The attachment power_saving_mode_spec.png is no longer available

Attached both binary and (Open)Watcom source (10KB binary for running what is basically two CPU instructions, sorry).
For using the binary , run "pwrcr on" in real mode (this means, no EMM386 loaded).
Me I'm getting about 100mA less on idle when on a HLT loop (so on DOS you need something like POWER.EXE loaded) .
Disclaimer: I have no idea if this will cause the CPU to catch fire with either short-term or longer-term use. This is just for experimenting and MOST LIKELY will have zero impact on your battery life (unless you leave the device idling/HLTd for significant periods of time).

Hmmm does that have any chance of being turned into a specific speed tuner for this SOC or may as well just use the classic ones? (AKA slowdown util like ATspeed, moslow etc to achieve XT compatibility speeds)

Anyone got the ps/2 mouse working on the BOOK386? Tested all ports (ps/2 keyboard, vga out, serial, parallel, game, usb, network) and all worked without problems. Serial mouse and touchpad are working but ps/2 mouse is not. Tried all combinations I could think off. First disabled only the touchpad (as described in the manual). Also tried disabling the mouse support in the bios and also disabling COM2. But none of the multiple mouse drivers I tried can find a ps/2 mouse.

I've had my Book386 for a couple of weeks now and while I haven't had much time to play with it, I do have a few thoughts:

• While it's bigger and heavier, the new design (vs the Pocket386) makes the system much more easily moddable. The internal expansion port on the Pocket386 uses a hard to source B2B connector where pretty much every connection internal to the Book386 is a simple 2.0mm pin header.
• Those internal headers also now break out pretty much everything. This was a big problem for the Pocket386 which broke out only a limited subset of the ISA bus and a set of pins on the back of the machine and lacked (for example) key DMA pins. So if you're like me and might be tempted to spin up an internal picogus that is now significantly less painful as a proposition than it was under the Pocket386 for both board area and connectors.
• As is now tradition one of the first things I did was replace the barrel jack with a USB-C trigger board wired for 12V. Not only are these now available pretty tiny but given the aforementioned spaciousness you can fit it in without even having the port lip poke out.
• The 2S 18650 lipo arrangement is interesting, and the battery compartment feels almost as if it's crying out for a second pair of cells (though it isn't quite wide enough).
• The GD5420 board is basically the same design as the one found in post Pocket386 orders, albeit also moving the LCD driver circuit from the screen in the Pocket386 and Hand386 to sit alongside. Slightly hobbled by the 512KB DRAM, I still had no trouble swapping the 5420 out for a 5428 for access to the blitting engine and theoretical access to greater colour depth (you just need to remove the strapping resistor at R16).
• As others have said, the inclusion of a coin cell is welcome (and confirmed, though I haven't checked its voltage) and I have high hopes this stops loss of RTC and BIOS config but also as others have already noted I've had at least one loss of BIOS config already so it's not perfect.